<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Simon Hardy on Medium]]></title>
        <description><![CDATA[Stories by Simon Hardy on Medium]]></description>
        <link>https://medium.com/@mesimonhardy?source=rss-5779be773682------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*utyDJKtDfJDMgMYpTG1cuQ.jpeg</url>
            <title>Stories by Simon Hardy on Medium</title>
            <link>https://medium.com/@mesimonhardy?source=rss-5779be773682------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sat, 21 Nov 2026 18:25:31 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@mesimonhardy/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Ontology, Hauntology, Late Capitalism and AI Rail-Guards: The Hauntological Continuum of the Status…]]></title>
            <link>https://ai.plainenglish.io/ontology-hauntology-late-capitalism-and-ai-rail-guards-the-hauntological-continuum-of-the-status-b7eae7b0469c?source=rss-5779be773682------2</link>
            <guid isPermaLink="false">https://medium.com/p/b7eae7b0469c</guid>
            <category><![CDATA[ontology]]></category>
            <category><![CDATA[philosophy]]></category>
            <category><![CDATA[ai-ethics]]></category>
            <category><![CDATA[democracy]]></category>
            <category><![CDATA[ai]]></category>
            <dc:creator><![CDATA[Simon Hardy]]></dc:creator>
            <pubDate>Mon, 24 Nov 2025 13:06:39 GMT</pubDate>
            <atom:updated>2025-11-24T13:06:39.254Z</atom:updated>
            <content:encoded><![CDATA[<h3><em>Ontology, Hauntology, Late Capitalism and AI Rail-Guards: The Hauntological Continuum of the Status Quo</em></h3><p>We have almost everything we need to redirect our technological trajectory, yet we remain trapped in a continuum of the status quo.</p><p>Hauntology and late capitalism explain why AI development keeps replaying and amplifying the past instead of opening real futures — <strong>while dynamic </strong><a href="https://en.wikipedia.org/wiki/Ontology"><strong>ontology</strong></a><strong>, combined with democratic rail-guards</strong>, shows how we might change course.</p><p>By “<a href="https://en.wikipedia.org/wiki/Hauntology"><strong>hauntological</strong></a><strong> continuum of the status quo</strong>”, I mean a form of historical time in which change is constant at the surface level (new products, new AIs, new crises), yet the underlying social, economic, and political structures remain fundamentally the same.</p><p>On this backdrop, we enter a new phase I call <strong>Das KI-Gespenst des Sozialstaatsendes</strong> or <strong>Hauntologische KI-Sozialstaatsdämmerung</strong>: <strong>the AI-ghost that haunts, signals, and normalizes the end of the welfare state</strong>; <strong>the hauntological twilight in which AI mediates and manages its gradual fading</strong>. In practice, this ghost functions as a systemic excuse — a softly lit, data-driven open jail.</p><p>By “<strong>AI hauntology of the end of the </strong><a href="https://en.wikipedia.org/wiki/Welfare_state"><strong>welfare state</strong></a>”, I mean the configuration in which AI no longer appears as a neutral innovation, but as a technology that manages, naturalizes, and aesthetically accompanies the erosion of the welfare state — a mode of futurity that allows only variations of the same dismantling.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Zxce0semX0_mr9R99L6Oaw.jpeg" /><figcaption>Photo by <a href="https://unsplash.com/@crinaparasca?utm_source=unsplash&amp;utm_medium=referral&amp;utm_content=creditCopyText">Crina Parasca</a> on <a href="https://unsplash.com/photos/black-and-pink-floral-textile-7HwPWVRA_qY?utm_source=unsplash&amp;utm_medium=referral&amp;utm_content=creditCopyText">Unsplash</a></figcaption></figure><h3>1. Introduction — When the Future Becomes a Ghost</h3><p>In public discourse, AI is almost always framed as “the future”: the next revolution, the next wave, the next inevitability. Yet the more AI systems spread through finance, policing, health care, and welfare administration, the more they feel strangely familiar. They do not open unknown worlds; <strong>they automate the ones we already know</strong> — with all their inequalities, exclusions, and blind spots—the future, <strong>promised as disruption</strong>, <strong>returns as a ghost with a diminished essence.</strong></p><p><strong>1.1 The central question</strong></p><blockquote><strong><em>Why does AI, a “future” technology, feel so trapped in old patterns of power, inequality, and extraction — and why is it increasingly mobilized to justify the dismantling of the welfare state rather than its renewal?</em></strong></blockquote><p>Behind the marketing language of efficiency and innovation, AI often appears as a tool for:</p><ul><li><strong>Tightening</strong> labour control rather than empowering workers;</li><li><strong>Targeting</strong> benefits rather than strengthening social rights;</li><li><strong>Optimizing</strong> austerity rather than expanding solidarity.</li></ul><p>Instead of designing new social futures, we are perfecting the <strong>administration of decline</strong>.</p><p><strong>1.2 From “innovation” to spectral repetition</strong></p><p>To make sense of this, we need to shift our lens from innovation to <strong>repetition</strong>. AI, in its dominant late-capitalist form, is not a neutral leap forward. It is the latest expression of a more profound/entangled/deeper logic:</p><ul><li>Relentless <strong>acceleration</strong> without structural transformation;</li><li>Endless <strong>novelty</strong> at the level of interfaces, apps, and products;</li><li>Near-total <strong>stasis</strong> at the level of property relations, power and institutional design.</li></ul><p>Data-driven systems trained on historical records encode exactly this pattern. They promise personalization while reproducing existing hierarchies; they promise prediction while extending the same risk logics that produced our current crises. AI becomes the technical arm of what <a href="https://dn790002.ca.archive.org/0/items/capitalist-realism-is-there-no-alternative/Capitalist%20Realism_%20Is%20there%20no%20alternativ%20-%20Mark%20Fisher.pdf">Mark Fisher called <strong>capitalist realism</strong>: the sense that “t<strong>here is no alternative</strong>”, <strong>only updates</strong>.</a></p><p><strong>1.3 Method and lenses</strong></p><p>To unpack this configuration — what I call the <strong>hauntological continuum of the status quo</strong> and, more specifically, the <em>AI hauntology of the end of the welfare state</em> — the article weaves together three main lenses:</p><ul><li><strong>Hauntology (</strong><a href="https://en.wikipedia.org/wiki/Jacques_Derrida"><strong>Derrida</strong></a><strong> / </strong><a href="https://en.wikipedia.org/wiki/Mark_Fisher"><strong>Mark Fisher</strong></a><strong>): </strong>Hauntology names a mode of time in which lost futures continue to haunt the present. It helps us understand why AI appears as a promise of the new while, in practice, it recycles and intensifies the old.</li><li><a href="https://en.wikipedia.org/wiki/Late_capitalism"><strong>Late capitalism</strong></a><strong> (Marxist / critical theory traditions): </strong>Late capitalism provides the structural backdrop: financialization, platformization, the culture industry, austerity, and the slow erosion of social protections. Critical theory shows how AI is embedded in this regime rather than standing outside it.</li><li><strong>Dynamic ontology (being as open and relational): </strong>Against static, essentialist views of society and technology, a dynamic ontology sees humans, institutions, and tools as co-evolving. This lens allows us to ask: <strong>what else could AI be if designed within a democratic, welfare-oriented framework rather than</strong> a purely extractive one?</li></ul><p>In the remainder of the article, these lenses converge to make a simple claim: <strong>AI is not destiny. It is a contested terrain on which the future of the welfare state </strong>— and the very possibility of alternatives to late capitalism — is <strong>being negotiated</strong>. To keep that future open, we will need not only technical safeguards but <strong>strong, explicitly democratic rail-guards</strong> that acknowledge both the hauntological weight of the past and the structural limits of algorithmic control.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*rZqtqQDBaOimeC0wHaJTJg.jpeg" /><figcaption>Photo by <a href="https://unsplash.com/@gnikomedi?utm_source=unsplash&amp;utm_medium=referral&amp;utm_content=creditCopyText">Gizem Nikomedi</a> on <a href="https://unsplash.com/photos/a-street-sign-on-the-side-of-a-road-PKjNJQDE0YA?utm_source=unsplash&amp;utm_medium=referral&amp;utm_content=creditCopyText">Unsplash</a></figcaption></figure><h3>2. Hauntology: The Future That Never Arrived</h3><p><strong>2.1 Hauntology: from Derrida to Mark Fisher</strong></p><p>Hauntology begins with a simple, unsettling intuition: <strong>the present is not fully present.</strong></p><p>When Jacques Derrida introduces the term in <em>Spectres of Marx</em>, he is trying to describe a strange feeling at the end of the Cold War. Officially, history is supposed to have “ended.” Capitalism has “won.” Yet the world is still full of <strong>ghosts</strong>: <strong>unfulfilled promises</strong>, <strong>struggles that were never resolved</strong>, <strong>futures that were declared dead but refuse to disappear</strong>.</p><p>For Derrida, hauntology is a way of talking about this paradoxical time:</p><ul><li>We are told that one system is final and inevitable;</li><li>Yet we are haunted by the sense that something else <strong>should have been possible</strong>.</li></ul><p>Mark Fisher later takes this intuition and drags it into everyday culture. For Fisher, hauntology is not just a philosophical mood; it becomes the <strong>structure of late-capitalist time itself</strong>. We live among ruins of abandoned futures:</p><ul><li>The political dreams of the 20th century;</li><li>The techno-utopian promises of early cyber-culture;</li><li>The social-democratic horizon of a steadily improving welfare state.</li></ul><p>These futures did not simply fail; they have been <strong>actively discredited</strong>, mocked, or buried under the claim that “<strong>there is no alternative.</strong>” And yet they keep returning as discomfort, nostalgia, or a diffuse sense of wrongness.</p><p>Hauntology, in this sense, names a world in which:</p><blockquote><em>The past is not past, the future has been cancelled, and the present is crowded with things that no longer properly belong to it.</em></blockquote><p><strong>It is this strange temporal knot that sets the stage for how AI is imagined, built, and deployed</strong>.</p><p><strong>2.2 Capitalist realism and the disappearance of alternatives</strong></p><p>Fisher’s key claim is brutally simple:</p><blockquote><em>“Under late capitalism, </em><strong><em>it becomes easier to imagine the end of the world than the end of capitalism</em></strong><em>”.</em></blockquote><p>This is what he calls <strong>capitalist realism</strong>: not a belief that capitalism is good or just, but a more profound, more corrosive conviction that it is <strong>the only viable system</strong>, the only game in town. Other ways of organizing society are not refuted; they are rendered <strong>unthinkable</strong>, unrealistic, childish.</p><p>Capitalist realism operates on several levels at once:</p><ul><li><strong>Economic:</strong> crises are treated as weather events; the basic structure of ownership and profit is never up for discussion.</li><li><strong>Political:</strong> elections become management changes, not horizon changes. Parties compete on efficiency within the same framework.</li><li><strong>Psychological:</strong> people internalize the message that “nothing can fundamentally change”; hope is privatized into self-help and individual coping strategies.</li></ul><p>This is where hauntology and capitalist realism converge. If there is no alternative, then all “new” technologies will be deployed <strong>inside the same framework</strong>:</p><ul><li>AI will be used to optimize markets, not question them;</li><li>Algorithms will target welfare fraud, not tax evasion by the richest;</li><li>Predictive systems will manage risk, not the causes of precarity.</li></ul><p>The result is a peculiar temporality:</p><ul><li>We are flooded with <strong>novelty</strong> (new versions, new apps, new models);</li><li>But all this novelty is steered toward <strong>maintaining the same basic order</strong>.</li></ul><p>That is the hauntological continuum of the status quo: the system moves, but it does not travel.</p><p><strong>2.3 Cultural hauntology: aesthetics of lost futures</strong></p><p>Hauntology is not only a matter of theory; it is <strong>audible and visible</strong> in culture. Fisher is particularly interested in music, but the pattern appears across media.</p><p>In music, hauntology often takes the form of:</p><ul><li>Samples of <strong>old public information films;</strong></li><li>Degraded tapes, vinyl crackle, retro synths;</li><li>Melodies that sound like half-remembered childhood jingles.</li></ul><p>It is not just nostalgia. The point is not “wasn’t the past nice?” but rather:</p><blockquote><em>“At some point, we genuinely believed the future would look different — and that future has vanished.”</em></blockquote><p>In film and television, we see:</p><ul><li>Endless reboots, remakes, and franchises;</li><li>“Retro-futuristic” aesthetics that recycle visions of tomorrow from the 60s, 70s, or 80s;</li><li>Dystopias that imagine catastrophe but rarely imagine <strong>different social relations</strong>.</li></ul><p>We are excellent at imagining (and framing a performative narrative that could trap us):</p><ul><li>Apocalypse;</li><li>Zombies;</li><li>Climate collapse,;</li><li>Corporate AI overlords…</li></ul><p>…but remarkably bad at imagining robust, credible alternatives to our current economic order.</p><p>This matters politically because culture is one of the main places where <strong>collective imagination is trained</strong>. If almost every story about the future presupposes the permanence of markets, platforms, and precarious labour, then it becomes harder and harder to think, let alone organize, for anything else.</p><p>Cultural hauntology, in other words, is not a quirky artistic niche; it is a <strong>symptom</strong> of capitalist realism. The lost futures we hear in music or see in film are the futures that have become structurally impossible to pursue under current conditions.</p><p><strong>2.4 Hauntology as a diagnostic tool for technology</strong></p><p>So what happens if we turn this lens onto AI?</p><p>The dominant story says:</p><ul><li>AI is a neutral tool;</li><li>It will “revolutionize everything”;</li><li>And its direction is mainly determined by technical feasibility/possibility.</li></ul><p>A hauntological reading suggests almost the opposite. AI does not appear in a vacuum; it is built, funded, and deployed within the <strong>hauntological continuum of late capitalism</strong>. That has at least four consequences.</p><p><strong>First, AI is trained on the past: </strong>Machine learning systems infer patterns from historical data. Suppose the past is structured by racism, class hierarchy, gendered division of labour, ableism, and colonial extraction. In that case, the default operation of AI is to <strong>encode and reproduce those structures exactly</strong>. Without deliberate counter-measures, AI systems will automate the spectral order we already have;</p><p><strong>Second, AI is steered by existing power:</strong> Who owns the models? Who owns the data? Who defines the optimization targets? At the moment, the answers are:</p><ul><li>Large corporations;</li><li>Security agencies;</li><li>And, indirectly, financial markets.</li></ul><p>Under these conditions, “innovation” tends to mean <strong>streamlining extraction and control</strong>: algorithmic management of workers, risk-scoring of welfare recipients, predictive policing, behavioural nudging, targeted advertising. <strong>The technology may be cutting-edge; the social imagination behind it is profoundly conservative.</strong></p><p><strong>Third, AI aesthetics participate in the ghosting of alternatives: </strong>Think of the marketing imagery: smooth interfaces, glowing lines of code, abstract “digital brains,” all floating in a frictionless blue. It suggests inevitability, neutrality, transcendence. What disappears from view are:</p><ul><li>Labour conditions in data centres and click-farms;</li><li>Ecological costs of hardware and energy,</li><li>The very real political choices about how and where AI is used.</li></ul><p>The result is that AI appears as a <strong>natural evolution rather than</strong> a c<strong>ontested socio-technical project</strong>. Its path seems as unquestionable as gravity.</p><p><strong>Fourth, AI becomes the perfect alibi for the end of the welfare state: </strong>In a hauntological welfare regime, cuts and restrictions are framed as <strong>technical necessities</strong>:</p><ul><li>“The algorithm has identified ineligible claimants.”;</li><li>“Data shows this hospital must close.”;</li><li>“Risk models require greater conditionality.”</li></ul><p>AI, presented as objective and impartial, serves to <strong>naturalize austerity</strong>. It becomes the polite, data-driven ghost that whispers: “<strong>There is no alternative to shrinking social protection; the system has spoken.</strong>”</p><p>In this way, AI is not simply a technology of the future. It is a machine that <strong>stabilizes a past-oriented order</strong>:</p><ul><li>It predicts based on yesterday;</li><li>Optimizes for the interests of today’s power-holders;</li><li>And quietly forecloses the futures that do not fit those trajectories.</li></ul><p>To say that AI is hauntological, then, is to say that it is entangled with a regime in which:</p><blockquote><em>The future is not genuinely open, but pre-formatted as more of the same — faster, more automated, more tightly managed.</em></blockquote><p>Hauntology, used as a diagnostic tool, allows us to see through the myth of AI as neutral progress. It reveals AI, in its current configuration, as one of the key <strong>operators of the hauntological continuum of the status quo</strong> — and, in the next step of the argument, <strong>as the spectral engineer of the welfare state’s dismantling</strong>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*WUyK8mNATscTaBixpo2ttw.jpeg" /><figcaption>Photo by <a href="https://unsplash.com/@alpridephoto?utm_source=unsplash&amp;utm_medium=referral&amp;utm_content=creditCopyText">Андрей Сизов</a> on <a href="https://unsplash.com/photos/a-book-with-a-diagram-on-it-nuz3rK5iiKg?utm_source=unsplash&amp;utm_medium=referral&amp;utm_content=creditCopyText">Unsplash</a></figcaption></figure><h3>3. Late Capitalism as a Hauntological Regime</h3><p>If hauntology gives us the mood — futures that never arrived, ghosts of alternative orders — <strong>late capitalism</strong> <strong>provides us with the machinery</strong>. It is not just one economic system among others; it’s a <strong>whole configuration</strong> of finance, technology, culture, and governance that produces precisely the kind of time we described:</p><ul><li>Endless movement;</li><li>Almost no transformation;</li><li>And a constant sense that “nothing else is realistically possible.”</li></ul><p>In that sense, late capitalism is not just the backdrop of hauntology; it is its <strong>engine</strong>.</p><p><strong>3.1 Structural features of late capitalism</strong></p><p>Late capitalism is often described through a list of features that sound abstract until you realize they map directly onto everyday life:</p><ul><li><strong>Financialization: </strong>Profits are increasingly made not by producing goods or providing services, but by trading, speculating, leveraging, and packaging risk. Pension funds, mortgages, student loans, sovereign debt — all become raw material for financial products. Decision-making across society is subordinated to “market confidence,” credit ratings, and shareholder value.</li><li><strong>Platformization: </strong>Large digital platforms position themselves as intermediaries for everything: communication, work, dating, transport, shopping, and entertainment. They operate as infrastructural monopolies: you don’t just use them, you live <em>inside</em> them. Their power lies not only in scale but also in their ability to set the rules for visibility, access, and monetization.</li><li><strong>Data extraction: </strong>Human experiences — clicks, movements, heartbeats, facial expressions, social connections — are turned into data for prediction and control. What used to be private or ephemeral becomes logged, mined, and modelled. “If it moves, measure it; if it can be measured, monetize it.”</li><li><strong>Precarity: </strong>Stable, long-term employment is eroded. People patch together incomes from gigs, short contracts, and “self-entrepreneurship.” The risk of economic breakdown — illness, unemployment, rent hikes — is displaced from the collective to the individual. Security becomes a personal project, not a social guarantee.</li><li><strong>The “always-on” economy: </strong>Work hours bleed into evenings and weekends. Devices keep us permanently reachable, trackable, and nudgable. Attention is commodified; rest is invaded. We are expected to be continuously responsive — to emails, to markets, to news cycles, to notifications.</li></ul><p>Taken together, these features generate a world that is:</p><ul><li><strong>Hyper-dynamic</strong> on the surface (prices, feeds, trends, rankings, updates);</li><li>But <strong>deeply conservative</strong> at the structural level (ownership, power, and extraction remain untouched).</li></ul><p>This is precisely what we earlier called the <strong>hauntological continuum of the status quo</strong>: a system that spins faster and faster to stay fundamentally the same.</p><p>AI fits into this not as an alien intrusion, but as a <strong>natural amplifier</strong>:</p><ul><li>Financialization uses AI for high-frequency trading and risk modelling;</li><li>Platforms use AI for recommendation, targeting, moderation, and ranking;</li><li>Data extraction is meaningless without machine learning systems to process the flood;</li><li>Precarious workers are managed, rated, and disciplined by algorithmic systems;</li><li>Predictive and nudging technologies orchestrate the always-on economy.</li></ul><p><strong>Late capitalism</strong> doesn’t just adopt AI; it <strong>shapes what AI is for</strong>.</p><p><strong>3.2 The culture industry and managed consciousness</strong></p><p>Adorno and Horkheimer, writing in the mid-20th century, spoke of the <strong>culture industry</strong>: film, radio, popular music, and advertising that standardized cultural products and, in doing so, standardized consciousness. The point was not that people were “brainwashed,” but that their <strong>imagination and desires</strong> were gently canalized:</p><ul><li>The same story structures;</li><li>The same aspirations;</li><li>The same norms, recycled in different packaging.</li></ul><p>Fast-forward to today, and the culture industry has gone <strong>interactive and algorithmic</strong>:</p><ul><li>Instead of a few national broadcasters, we have global platforms (TikTok, YouTube, Instagram, Spotify, Netflix, etc.);</li><li>Instead of one program schedule for everyone, we have <strong>personalized feeds</strong> curated by recommender systems;</li><li>Instead of passive spectatorship, we have constant micro-participation: likes, comments, shares, swipes, posts.</li></ul><p>It looks and feels more democratic — “you choose what you like” — but the choices <strong>are pre-filtered by opaque systems optimizing for engagement, advertising revenue, and platform growth</strong>.</p><p>Managed consciousness 2.0 works through:</p><ul><li><strong>Algorithmic visibility:</strong> what you see is what the system predicts will keep you there; what you <em>don’t</em> see quietly ceases to exist for you;</li><li><strong>Feedback loops:</strong> your clicks train the model, the model shapes what you see, what you see shapes your beliefs and habits, which in turn train the model further;</li><li><strong>Emotional calibration:</strong> outrage, anxiety, envy, and fear are often more “engaging” than calm, reflective content — so the system tilts in that direction.</li></ul><p>In this sense, contemporary culture helps to <strong>normalize</strong> the structural features of late capitalism:</p><ul><li>Hustle culture and “personal brands” make precarity feel like a lifestyle choice;</li><li>Techno-solutionist narratives present platforms and AI as the only realistic tools for solving social problems;</li><li>Dystopias entertain us with catastrophe while leaving the basic structure — corporations, states, markets — weirdly intact.</li></ul><p>What disappears is the <strong>thinkability of alternatives</strong>:</p><ul><li>Public ownership of platforms;</li><li>Decommodified housing;</li><li>Guaranteed income;</li><li>Deep democratization of workplaces and institutions.</li></ul><p>The imaginal space is managed, and AI-powered recommendation systems are now central devices of that management.</p><p><strong>3.3 Systemic resistance to change</strong></p><p>Given all this, it is not surprising that late capitalism is <strong>structurally allergic to deep reform</strong>. It can tolerate crisis, scandal, and even partial regulation, but it tends to deflect anything that might alter its core logic.</p><p>There are several reasons:</p><p><strong>Profit dependence on extraction: </strong>When business models depend on capturing attention, monetizing data, and squeezing labour costs, any reform that threatens those extraction channels appears as an existential threat. “Reforms” are accepted only when they can be subsumed into new profit strategies.</p><p><strong>Short-termism: </strong>Financial markets reward short-term gains. Companies are under constant pressure to show quarterly growth. Long-term investments in social or ecological resilience look irrational within this framework, even if they are rational for society as a whole.</p><p><strong>Path dependence and lock-in: </strong>Existing infrastructures, legal frameworks, business practices, and everyday habits form a tight mesh. Changing one part without changing the rest is hard. Even when alternatives exist, the cost of switching — politically, economically, cognitively — is high.</p><p><strong>Risk externalization: </strong>The system is sound by design in externalizing risk: ecological damage, mental health burdens, social fragmentation, and democratic erosion are treated as unfortunate but acceptable side effects. If decision-makers do not bear the costs of non-reform, there is little incentive for profound change.</p><p>In this context, <strong>meaningful AI regulation</strong> faces predictable resistance:</p><ul><li>Strong data protection rules limit extraction;</li><li>Algorithmic transparency undermines trade secrets and proprietary advantages;</li><li>Restrictions on certain uses (mass surveillance, predictive policing, welfare scoring) clash with the interests of both states and corporations;</li><li>Labour protections and union oversight challenge the use of AI as a disciplinary tool.</li></ul><p>So we get instead:</p><ul><li>Voluntary ethics guidelines;</li><li>Advisory boards with limited power;</li><li>Soft law and “best practices”, while crucial questions — ownership, business models, enforcement — remain unchanged.</li></ul><p>This is hauntological in an exact sense:</p><blockquote><em>What is proposed as “new” governance reproduces the same basic pattern — </em><strong><em>the system performs reform while essentially immunizing itself against transformation</em></strong><em>.</em></blockquote><p>AI, in this story, is framed as inevitable and unstoppable, which <strong>conveniently justifies incremental, cautious, technocratic responses</strong> rather than bold political decisions.</p><p><strong>3.4 Political interests and captured institutions</strong></p><p>Finally, late capitalism is not just an abstract system; it is maintained and steered by <strong>concrete political interests</strong> operating through <strong>very real institutions</strong>.</p><p><strong>Lobbying and influence networks: </strong>Large tech companies and financial actors invest heavily in lobbying, consultancy, think tanks, and public-private partnerships. They frame debates, draft legislative proposals, sit on expert panels, and supply “technical expertise” that often aligns with their own interests.</p><p><strong>Revolving doors: </strong>People move between government, regulatory bodies, and corporate leadership. This creates shared worldviews and priorities: innovation, competitiveness, “cutting red tape,” and partnership with the private sector. Structural critique of capitalism or serious expansion of welfare protections rarely fits this horizon.</p><p><strong>Regulatory capture: </strong>Institutions meant to regulate industries increasingly rely on those same industries for information, tools, and cooperation. Over time, regulation becomes shaped by what is acceptable to the regulated. In AI governance, this can mean:</p><ul><li>Standards written with a strong corporate imprint;</li><li>Oversight agencies are underfunded and understaffed;</li><li>Enforcement that targets small actors but rarely challenges systemic abuses of major players.</li></ul><p><strong>Fiscal and technological dependence of states: </strong>States depend on global capital flows, tax revenues, and access to cutting-edge technologies. Governments that threaten major corporate interests fear capital flight, investment strikes, <strong>or being “left behind” in the global AI race</strong>. This dependence narrows the space of politically “realistic” options.</p><p>The consequence is that the <strong>spectral status quo</strong> you already named:</p><ul><li>Policies that appear new but leave the underlying power structure intact;</li><li>AI strategies that promise “trustworthy” or “human-centric” systems while deepening surveillance, precarity, and welfare retrenchment;</li><li>Continuous consultation processes that produce frameworks, roadmaps, and principles — but little that materially changes who owns what, who decides what, and who is protected from what.</li></ul><p>In such a landscape, AI is quickly woven into the existing institutional fabric as:</p><ul><li>A tool for “evidence-based” austerity;</li><li>A justification for automation and workforce reduction;</li><li>An instrument for risk scoring and behavioral management;</li><li>All under the banner of modernization.</li></ul><p>This is what it means to say that <strong>late capitalism functions as a hauntological regime</strong>:</p><ul><li>It organizes time so that the past (its own logic of extraction and hierarchy) dominates the future;</li><li>It deploys culture and technology to manage imagination and behaviour;</li><li>It adapts to shocks and critiques without easily conceding structural change.</li></ul><p>AI, under these conditions, is not merely a neutral innovation. It becomes one of the central <strong>operators of this regime</strong> — a spectral administrator of the present order, <strong>a key device through which the end of the welfare state can be presented as a technical necessity rather than a political choice</strong>.</p><p>In the next step of the argument, this is precisely why <strong>dynamic ontology</strong> and <strong>strong, democratic rail-guards</strong> become crucial: not as decorative ethical add-ons, but as deliberate interruptions of a system that would otherwise use AI to perfect the very status quo that is driving social and ecological breakdown.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*4FC2TK0AowOvkTk6O33E_Q.jpeg" /><figcaption>Photo by <a href="https://unsplash.com/@guyyama?utm_source=unsplash&amp;utm_medium=referral&amp;utm_content=creditCopyText">Guy Yama</a> on <a href="https://unsplash.com/photos/a-close-up-of-a-planet-K2RDHAfkLu8?utm_source=unsplash&amp;utm_medium=referral&amp;utm_content=creditCopyText">Unsplash</a></figcaption></figure><h3>4. Dynamic Ontology vs Essential Ontology</h3><p>Underneath our debates about AI, welfare, and democracy sits a quieter question that usually goes unnamed:</p><blockquote><strong><em>What do we think “being” is like?</em></strong></blockquote><p>Are humans, institutions, and technologies basically <strong>fixed things</strong> that reveal a pre-given essence?</p><p>Or are they <strong>unfinished processes</strong>, constantly becoming and reconfiguring in relation to others?</p><p>This is where the contrast between <strong>essential ontology</strong> and <strong>dynamic ontology</strong> becomes decisive.</p><p><strong>4.1 Essential ontology: being as fixed identity</strong></p><p>Essential ontology treats things — especially people and social roles — as if they have a <strong>stable, inner nature</strong> that defines them once and for all.</p><p>It shows up in everyday language as:</p><ul><li>“That’s just human nature.”</li><li>“People like that will always…”</li><li>“The market is naturally efficient.”</li><li>“This group is inherently more… [rational/emotional/violent/peaceful].”</li></ul><p>In political and social terms, essential ontology tends to produce:</p><p><strong>Static categories:</strong></p><ul><li>“deserving” vs “undeserving” poor;</li><li>“high-skilled” vs “low-skilled” workers;</li><li>“risky” vs “trustworthy” populations.</li></ul><p>These categories are treated as if they reflect <strong>profound truths</strong> rather than context, history, and power.</p><p><strong>Rigid hierarchies: </strong>If some people are “naturally” more rational, disciplined, or productive, then it seems reasonable that they should rule, earn more, or be less constrained by rules. Inequality appears as an expression of essence, not of injustice.</p><p><strong>Moral alibis: </strong>Essentialism turns structural problems into natural facts:</p><ul><li>Poverty as a result of “poor choices”;</li><li>Crime is “inherent” to specific communities;</li><li>Inequality as the inevitable outcome of talent and effort.</li></ul><p>In this framework, <strong>change</strong> is either cosmetic (better management, better incentives) or framed as dangerous: anything that really shifts structures is seen as an attack on order itself.</p><p>Essential ontology is therefore very comfortable for a hauntological regime:</p><ul><li>If the fundamental nature of things is fixed;</li><li>And the institutions built on that nature are legitimate;</li><li>Then the only “realistic” future is <strong>more of the same</strong>, perhaps with “smarter”/more innovative tools.</li></ul><p>AI fits perfectly into this story when it is sold as a way to <strong>reveal the underlying truth</strong> of people and situations — who will repay a loan, who is a risk, who is productive — by reading the “data” of their essence.</p><p><strong>4.2 Dynamic ontology: being as becoming &amp; relation</strong></p><p>Dynamic ontology starts somewhere very different. Instead of treating beings as finished objects with fixed traits, it sees them as <strong>processes in relation</strong>:</p><blockquote><em>Being is not a static state but an ongoing </em>becoming<em> — co-created through relationships, history, and choice.</em></blockquote><p>In this view:</p><ul><li>A person is not just “what they are” but <strong>who they are becoming over time</strong>, through encounters, learning, trauma, healing, responsibility, and response;</li><li>Institutions are not timeless structures; they are <strong>crystallized histories</strong> that can be reformed, repented of, or rebuilt;</li><li>Technologies are not neutral tools; they are <strong>participants</strong> in social becoming, shaped by values and shaping them in turn.</li></ul><p>Dynamic ontology creates room for concepts like:</p><ul><li><strong>Reorientation/Recalibration</strong> — the possibility of return, turning, reorientation. People and societies can recognize harm, change course, and write new chapters;</li><li><strong>Social transformation</strong> — not just policy tweaks, but <strong>real reconfiguration</strong> of power, property, and protection;</li><li><strong>New institutional forms</strong> — cooperatives, commons, public platforms, citizen assemblies, regenerative economies — all of which assume that our current arrangements are not final.</li></ul><p>This perspective does not deny limits or continuity. It does not pretend that we can become anything, instantly, by wish. But it insists that:</p><ul><li>Identities are <strong>historical</strong>, not metaphysical;</li><li>Systems are <strong>path-dependent</strong>, not predestined;</li><li>And the future is <strong>open enough</strong> to justify responsibility, imagination, and struggle.</li></ul><p>Dynamic ontology is what makes hope rational rather than naïve.</p><p><strong>4.3 Why ontology matters for AI and democracy</strong></p><p>At first glance, this may sound abstract. But ontology quietly shapes <strong>how we design AI</strong> and <strong>how we justify our institutions</strong>.</p><p>If we lean toward <strong>essential ontology</strong>, we tend to:</p><ul><li>Treat data as a <strong>mirror of the true self</strong> or the true nature of groups;</li><li>Use AI to <strong>classify, sort, and fix</strong> people into stable risk categories, credit scores, productivity scores, and behavioral profiles;</li><li>Design systems that <strong>conserve the past</strong>: the best we can do is reveal patterns more accurately and administer existing hierarchies more efficiently.</li></ul><p>If we lean toward <strong>dynamic ontology</strong>, design looks very different:</p><ul><li>We see data as <strong>partial traces</strong> of situated lives, not as a complete record of essence;</li><li>We design AI systems that <strong>support repair and change</strong>: tools for identifying structural bias, mechanisms for helping people out of debt or precarity rather than locking them into “high-risk” labels, systems that can be appealed, contested, and revised through democratic processes.</li></ul><p><strong>For democracy, dynamic ontology means</strong>:</p><ul><li>Citizens are not fixed interest-bearers but <strong>agents capable of learning and solidarity</strong>;</li><li>Institutions are open to <strong>re-design</strong>, not treated as sacrosanct;</li><li>Collective decisions are about <strong>what kind of beings we are becoming together</strong>, not just how to manage what already exists.</li></ul><p>In short:</p><blockquote><strong><em>If we think beings are fixed, we build AI to </em>predict and police<em>.</em></strong></blockquote><blockquote><strong><em>If we think beings are capable of becoming, we build AI to </em>support and transform<em>.</em></strong></blockquote><p>Ontology is not a footnote. It’s the hidden blueprint behind technical architectures and legal frameworks.</p><p><strong>4.4 Dynamic ontology as an antidote to hauntology</strong></p><p>Hauntology told us that under late capitalism, the future is colonized by the past:</p><ul><li>Lost futures haunt us;</li><li>Alternatives feel impossible;</li><li>The present is stretched out into an endless sequel.</li></ul><p>Dynamic ontology is not a magic escape hatch, but it does provide a <strong>counter-move</strong>.</p><p>Where hauntology observes:</p><ul><li>Repetition of the same structures under new interfaces;</li><li>The cancellation of credible alternatives;</li><li>The ghostly return of disavowed hopes,</li></ul><p><strong>Dynamic ontology insists</strong>:</p><ul><li>Structures can be <strong>repented of and reworked</strong>;</li><li>Alternatives are not fantasies but <strong>unrealized branches</strong> of our shared story;</li><li>The future is not a pre-programmed endpoint but a <strong>field of contested possibilities</strong>.</li></ul><p>Applied to AI and the welfare state, this means:</p><ul><li>We refuse the narrative that the erosion of social protection is a “technical inevitability”;</li><li>We expose how AI is currently used to <strong>administer decline</strong> and <strong>normalize exclusion;</strong></li><li>And we reclaim AI as a potential <strong>tool-ladder</strong>: something that could, if governed differently, help us climb toward greater justice, care, and solidarity instead of descending into automated austerity.</li></ul><p>Dynamic ontology, then, does two crucial things in this article:</p><ol><li>It names the <strong>philosophical error</strong> behind many AI inevitability claims: the assumption that data reveals essences and that those essences justify hierarchy;</li><li>It underwrites the <strong>political possibility</strong> of strong, democratic rail-guards. If being is open and relational, then designing new institutional constraints and protections is not an act against “nature,” but an act in line with how reality actually works.</li></ol><p>From this angle, <strong>rail-guards</strong> are not obstacles to progress; they are <strong>expressions of a different ontology of progress</strong>: one where we do not optimize a haunted status quo, but deliberately open paths toward futures in which more people can live, act, and change with dignity.</p><p>In other words:</p><blockquote><em>Against a ghostly repetition of the same, dynamic ontology allows for genuine future-making — and AI, under that horizon, must be judged not by how well it predicts yesterday, but by how much it helps us build a more just tomorrow.</em></blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*2_4peK_nt7XTFXFBoHCAAQ.jpeg" /><figcaption>Photo by <a href="https://unsplash.com/@purzlbaum?utm_source=unsplash&amp;utm_medium=referral&amp;utm_content=creditCopyText">Claudio Schwarz</a> on <a href="https://unsplash.com/photos/the-word-justice-written-in-black-ink-on-a-concrete-surface-SdoPqamuCcw?utm_source=unsplash&amp;utm_medium=referral&amp;utm_content=creditCopyText">Unsplash</a></figcaption></figure><h3>5. The Frankfurt School, Democracy, and the Welfare State</h3><p>Suppose hauntology gives us a way to read time, and dynamic ontology gives us a way to think being. In that case, <strong>critical theory</strong> provides us with a way to read <strong>power</strong> — especially when it hides behind “progress,” “innovation,” or “modernization.”</p><p>The Frankfurt School isn’t a museum piece here; it’s a toolkit for decoding why the very institutions that could protect us — democracy, welfare, law — are being quietly hollowed out while AI is rolled in as the new organizing myth.</p><p><strong>5.1 Critical theory vs “neutral” progress narratives</strong></p><p>Mainstream narratives about technology and the economy like to sound <strong>neutral</strong>:</p><ul><li>“AI will increase productivity.”</li><li>“Digitalization is inevitable.”</li><li>“We must innovate to stay competitive.”</li></ul><p>These phrases are presented as simple facts, like the weather forecast. What Adorno, Horkheimer, Marcuse, Fromm, and others tell us is:</p><blockquote><em>Whenever a social order describes its own trajectory as neutral, inevitable, or “just how things are,” you are already inside </em><strong><em>ideology</em></strong><em>.</em></blockquote><p>For the Frankfurt School, <strong>reason</strong> has a dark side: it can degenerate into <strong>instrumental reason</strong> — <strong>intelligence aimed only at control, efficiency, and domination</strong>, detached from questions of justice, meaning, or truth.</p><p>Critical theory asks:</p><ul><li>Whose interests are served when we call something “progress”?</li><li>Who benefits from this particular “innovation”?</li><li>Which other futures are silently declared unrealistic or irrational?</li></ul><p>Applied to AI:</p><ul><li>“Optimization” for whom?</li><li>“Efficiency” at whose expense?</li><li>“Innovation” according to whose priorities and metrics?</li></ul><p>Critical theory trains us to see that:</p><ul><li>A “neutral” risk model can embody class and racial bias;</li><li>A “smart city” can mean intensified surveillance and exclusion;</li><li>A “modernized welfare system” can mean automated sanctions and stricter conditionality.</li></ul><p>The Frankfurt School doesn’t give us a blueprint for AI, but it gives us the <strong>habit of suspicion</strong> we desperately need: never take technical progress at face value; <strong>always ask how it fits into broader patterns of domination or emancipation</strong>.</p><p><strong>5.2 The welfare state as a fragile achievement</strong></p><p>From a critical theory perspective, the <strong>welfare state</strong> is not a natural or inevitable part of capitalism. It is:</p><blockquote><em>A contingent, hard-won interruption in pure market logic.</em></blockquote><p>Things like:</p><ul><li>Social Security;</li><li>Public health care;</li><li>Unemployment insurance;</li><li>Labour protections;</li><li>Pensions;</li><li>Public education…</li></ul><p>did not appear because markets “evolved” toward kindness. They were:</p><ul><li><strong>Fought for</strong> by workers’ movements, unions, and social struggles;</li><li><strong>Conceded</strong> under pressure, often to avoid more radical transformations;</li><li><strong>Institutionalized</strong> through laws and bureaucracies that had to be built and defended.</li></ul><p>The welfare state represents a fragile compromise:</p><ul><li>Capitalism remains;</li><li>But some of its harshest consequences are <strong>partially buffered</strong> by collective guarantees.</li></ul><p>Fromm’s work on freedom, alienation and the need for belonging helps us see that:</p><ul><li>The welfare state is not just a technical arrangement;</li><li>It is a <strong>moral and psychological achievement</strong>: a recognition that people should not be left alone with market risks, that dignity requires more than “opportunity”.</li></ul><p>Seen this way, the welfare state is not “old-fashioned bureaucracy” standing in the way of innovation. It is one of the few institutional structures that <strong>push back</strong> against the very forces late capitalism wants to universalize: competition, commodification, and individualization of risk.</p><p>This makes it a natural target for erosion.</p><p><strong>5.3 Erosion under late capitalism</strong></p><p>Under late capitalism, we have watched a slow, uneven but persistent <strong>hollowing-out</strong> of both democracy and the welfare state. The vocabulary of this erosion is familiar:</p><ul><li><strong>Austerity</strong> — framed as tough but necessary discipline after crises;</li><li><strong>Privatization</strong> — sold as efficiency and choice;</li><li><strong>New Public Management</strong> — importing corporate logic into public services;</li><li><strong>Activation policies</strong> — shifting responsibility onto individuals to be constantly employable.</li></ul><p>The consequences are equally familiar:</p><ul><li>Underfunded hospitals and schools;</li><li>Social services turned into bureaucratic obstacle courses;</li><li>Increased conditionality for benefits;</li><li>Precarious work framed as flexibility;</li><li>Chronic stress and burnout treated as individual failure or “resilience” challenge.</li></ul><p>Democratic structures suffer parallel trends:</p><ul><li>Parties converge on narrow economic orthodoxy;</li><li>Technocratic bodies and expert committees increasingly shape policy decisions;</li><li>Citizens are repositioned as consumers or data points rather than co-authors of the common good.</li></ul><p>In this context, AI enters public discourse as both <strong>solution</strong> and <strong>alibi</strong>:</p><ul><li>Algorithmic “fraud detection” to justify stricter welfare policing;</li><li>Predictive models to rationalize hospital closures or school mergers;</li><li>Automated “risk scoring” to target social interventions instead of expanding universal provision.</li></ul><p>The story is always the same:</p><blockquote><em>“We can’t afford the old model; the numbers don’t lie; the algorithm has shown us the way.”</em></blockquote><p>But behind this story lies a choice:</p><ul><li>Instead of <strong>reinvesting</strong> in welfare and democracy,</li><li>We are using technology to <strong>administer decline</strong> and <strong>distribute scarcity</strong> — often in ways that punish the already vulnerable.</li></ul><p>From a hauntological perspective, this is the <strong>AI-managed twilight of the welfare state</strong>: not a sudden collapse, but a long sunset where social protections dim while the language of modernization, optimization and innovation grows brighter.</p><p><strong>5.4 Why critical theory is essential for AI governance</strong></p><p>In many policy spaces, <strong>AI ethics</strong> appears as a set of principles:</p><ul><li>Fairness;</li><li>Transparency;</li><li>Accountability;</li><li>Human-centric design.</li></ul><p>These are not bad words. But without a critique of <strong>power</strong> and <strong>ideology</strong>, they risk becoming:</p><blockquote><em>A decorative layer on top of systems that remain fundamentally organized around extraction, control and inequality.</em></blockquote><p>This is where the Frankfurt School becomes essential, not optional.</p><p>Critical theory reminds us that:</p><ul><li>You cannot evaluate AI systems in isolation from <strong>who owns</strong> them, <strong>who benefits</strong> from them, and <strong>which social order</strong> they are embedded in;</li><li>“Bias” is not just a technical glitch; it is the <strong>reflection of structural injustice</strong> in data and models;</li><li>“Trustworthy AI” can itself become a <strong>branding exercise</strong> if it doesn’t confront underlying economic and political logics.</li></ul><p>For AI governance, this means:</p><ul><li>Ethics must be linked to <strong>material constraints</strong>: bans, limits, rights, enforcement, and redistribution — especially in welfare, policing, labour, and migration;</li><li>Governance has to include <strong>questions of ownership and control</strong>: should certain AI infrastructures be public, cooperative, or commons-based rather than corporate monopolies?</li><li>Democratic participation cannot be reduced to user feedback; it must involve <strong>citizen assemblies, unions, civil society</strong>, and those most affected by AI-driven decisions.</li></ul><p>Without this critical lens:</p><ul><li>AI regulation risks focusing on <strong>procedural fixes</strong> (impact assessments, checklists, audits) while leaving intact the use of AI as a tool for austerity and social control;</li><li>The welfare state can be quietly reprogrammed into a <strong>conditional, punitive apparatus</strong>, with algorithms providing the appearance of objectivity.</li></ul><p>Critical theory, combined with dynamic ontology, pushes us to a harder but more honest question:</p><blockquote><em>Are we building AI to </em><strong><em>serve</em></strong><em> a democratic, solidaristic welfare state,</em></blockquote><blockquote><em>or are we deploying AI to </em><strong><em>replace and discipline</em></strong><em> it?</em></blockquote><p>If we don’t ask that, “AI governance” will amount to fine-tuning the machinery of a hauntological regime: the future endlessly promised as innovation, while the actual horizon narrows.</p><p>To avoid that, we need governance that:</p><ul><li>Recognizes the <strong>fragility and value</strong> of the welfare state;</li><li>Sees <strong>through</strong> neutral progress narratives;</li><li>And dares to say that some uses, models and business strategies are <strong>incompatible</strong> with a just, democratic order — no matter how innovative they are.</li></ul><p><strong>Only then can AI rail-guards become what they must be in this article</strong>: not polite guidelines around a haunted status quo, but <strong>structural interventions</strong> that help re-open the future and protect the conditions for real human flourishing.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*fROu3w5qLYW9yB3K-fFIBA.jpeg" /><figcaption>Photo by <a href="https://unsplash.com/@choys_?utm_source=unsplash&amp;utm_medium=referral&amp;utm_content=creditCopyText">Conny Schneider</a> on <a href="https://unsplash.com/photos/a-3d-image-of-a-judges-hammer-on-a-black-background-3hkKv6WzjcE?utm_source=unsplash&amp;utm_medium=referral&amp;utm_content=creditCopyText">Unsplash</a></figcaption></figure><h3>6. AI as a Hauntological Machine</h3><p>By this point, AI no longer appears as an abstract “future technology,” but as something built <em>inside</em> a very specific order: late capitalism as a hauntological regime. Now we can say the quiet part out loud:</p><blockquote><em>In its dominant form, AI is not just a set of tools. It is a </em><strong><em>hauntological machine</em></strong><em>: a way of turning the frozen past into an apparently objective guide for the future.</em></blockquote><p>It does this through four main functions: how it learns, how it is deployed as infrastructure, how it shapes imagination, and how it externalizes its costs.</p><p><strong>6.1 Training on the past: when datasets become ghosts</strong></p><p>Most contemporary AI systems are trained on historical data:</p><ul><li>Credit histories;</li><li>Past hiring decisions;</li><li>Historical crime records;</li><li>Medical records;</li><li>Browsing behaviour;</li><li>Social media interactions,</li></ul><p>and so on.</p><p>This is usually framed as neutral: we “learn patterns” from reality. But the “reality” inside these datasets is already:</p><ul><li>Structured by <strong>racism, sexism, class hierarchy, ableism, and colonial relations;</strong></li><li>Filtered through <strong>institutional practices</strong> (who gets recorded, who gets policed, who gets tested);</li><li>Shaped by <strong>power</strong> (who can avoid surveillance, who can’t).</li></ul><p>When these datasets are treated as <strong>ground truth</strong>, three things happen:</p><p><strong>Historical bias becomes mathematical authority: </strong>Discriminatory patterns — in who gets loans, who gets hired, who gets stopped by police — are absorbed as if they were natural facts. The model is not told: “This is injustice.” It is told: “This is reality. Predict this.”</p><p><strong>Structural injustice becomes a prediction target: </strong>The purpose of many models is to guess:</p><ul><li>Who will default;</li><li>Who will re-offend;</li><li>Who will quit a job;</li><li>Who will need benefits.</li></ul><p>But if the system is trained on a world where some groups face more barriers and punishment, then <strong>risk</strong> is just a proxy for existing oppression.</p><p><strong>Power relations are frozen in code: </strong>“Ground truth” is often whatever existing institutions have decided:</p><ul><li>The judge’s ruling;</li><li>The bank’s acceptance or refusal;</li><li>The doctor’s diagnosis;</li><li>The welfare agency’s sanction.</li></ul><p>AI then learns to <strong>imitate those decisions</strong> at scale.</p><p>In hauntological terms, datasets become <strong>ghost archives</strong>: they memorialize a specific configuration of power and harm, and AI systems call those ghosts back, over and over, every time a prediction is made.</p><p>Instead of asking, “How can we change these patterns?”, the system asks, “How can we reproduce them more efficiently?”</p><p><strong>6.2 Platform capitalism and algorithmic governance</strong></p><p>AI doesn’t just live in research labs. It is embedded in the infrastructure of <strong>platform capitalism</strong>:</p><ul><li>Social media feeds are ordered by recommender systems;</li><li>Gig work is assigned, timed, and evaluated by algorithms;</li><li>Online markets and app ecosystems are curated by ranking and recommendation;</li><li>Search results and news exposure are filtered and prioritized by opaque models.</li></ul><p>This is <strong>algorithmic governance</strong>: rules and decisions about what is visible, who gets access, and who is rewarded are increasingly made by systems that:</p><ul><li>Monitor behaviour;</li><li>Predict responses;</li><li>Nudge choices;</li><li>And constantly update themselves.</li></ul><p>Several logics converge here:</p><p><strong>Surveillance: </strong>Platforms and institutions collect detailed behavioural traces — where you go, what you watch, how long you hover, what you buy, who you talk to. AI systems turn this into profiles and predictions.</p><p><strong>Nudging and micro-targeting: </strong>Instead of issuing clear commands, systems shape the environment:</p><ul><li>Personalized recommendations;</li><li>Targeted ads;</li><li>“You may also like”;</li><li>Subtle design choices that push you toward certain actions;</li><li>Power appears as friendly suggestion, not as order.</li></ul><p><strong>Labour control: </strong>Workers in warehouses, call centres, ride-hailing and delivery platforms, even professional jobs, are increasingly:</p><ul><li>Tracked in real time;</li><li>Scored on performance metrics;</li><li>Scheduled algorithmically.</li></ul><p>AI becomes a <strong>digital foreman</strong>, enforcing productivity quotas and penalizing deviation, often without a human supervisor who can be meaningfully appealed to.</p><p><strong>Infrastructural monopoly: </strong>Because these systems run on massive data and computing infrastructures, they tend to be centralized in a few corporate actors. Their models and metrics become quasi-public infrastructure without being publicly governed.</p><p>This is hauntological in a deep sense:</p><blockquote><em>The same logic — optimize engagement, extract data, discipline labour — runs underneath many different interfaces, quietly steering behaviour toward the maintenance of the existing order.</em></blockquote><p>AI here is not just a feature. It is the <strong>operating system</strong> of platform capitalism.</p><p><strong>6.3 The automation of capitalist realism</strong></p><p>Capitalist realism, as we saw, is the sense that “there is no alternative” to the current system. AI contributes to this not just ideologically, but <strong>operationally</strong>.</p><p>Several mechanisms are at work:</p><p><strong>Prediction as destiny: </strong>When models label someone “high risk,” “low value,” “low potential,” those labels shape:</p><ul><li>Whether they get a loan;</li><li>Whether they are hired;</li><li>Whether they are monitored more closely.</li></ul><p>The prediction helps <em>produce</em> the very outcome it forecasts. What was a probability becomes a self-fulfilling pattern.</p><p><strong>Optimization of the existing game: </strong>Most AI deployments are framed as:</p><ul><li>“Improving efficiency”;</li><li>“Reducing friction”;</li><li>“Maximizing engagement”;</li></ul><p>Within existing markets and institutions.</p><p>The question is rarely, “Is this game worth playing?” but “How do we play it better?” This narrows imagination to <em>tweaks rather than</em> transformations.</p><p><strong>Displacement of responsibility: </strong>Decisions can be attributed to:</p><ul><li>“What the model says”;</li><li>“What the data shows”;</li><li>“What the risk profile indicates.”</li></ul><p>Structural choices (like underfunding welfare, outsourcing work, intensifying policing) are framed as <strong>technical necessities</strong> rather than political acts.</p><p><strong>Narratives of inevitability: </strong>AI is presented as something we must “adapt to,” much like the weather or gravity. This undermines the sense that society could collectively choose:</p><ul><li>Different goals;</li><li>Different uses;</li><li>Or even prohibitions on certain applications.</li></ul><p>The overall effect is the <strong>automation of capitalist realism</strong>:</p><blockquote><em>The sense that “there is no alternative” is no longer just a feeling; it is baked into prediction systems, optimization targets, and governance procedures that constantly select for the continuity of the existing order.</em></blockquote><p>In this way, AI becomes a key operator of the <strong>hauntological continuum of the status quo</strong>: it turns yesterday’s power relations into metrics, feeds them into models, and uses them to shape tomorrow.</p><p><strong>6.4 Ecological and social externalities</strong></p><p>AI is often marketed as immaterial: “the cloud,” “digital infrastructure,” “intelligence.” In reality, it is rooted in very material processes that generate significant <strong>ecological</strong> and <strong>social</strong> externalities.</p><p><strong>Energy use and emissions: </strong>Training and running large models consume enormous amounts of electricity. Data centres rely on physical infrastructure, cooling systems, and energy grids. The environmental costs are often shifted onto:</p><ul><li>Local communities near data centres;</li><li>Regions where electricity is cheap (not always clean);</li><li>Future generations who inherit the climate burden.</li></ul><p><strong>Resource extraction: </strong>Hardware for AI — servers, GPUs, smartphones, cables — depends on:</p><ul><li>Mining of rare earths and metals;</li><li>Water use for cooling and processing;</li><li>Global supply chains with significant ecological footprints.</li></ul><p>Extraction sites are frequently in regions with weaker protections, where environmental damage and labour exploitation are easier to externalize.</p><p><strong>Labour exploitation in the AI supply chain: </strong>Behind the polished interface of “intelligent” systems lies:</p><ul><li>Clickworkers labelling data for cents per task;</li><li>Content moderators exposed to traumatic material;</li><li>Warehouse and delivery workers monitored by the very systems they help sustain.</li></ul><p>Much of this work is precarious, low-paid, and geographically displaced from the spaces where AI is celebrated as “innovation.”</p><p><strong>New forms of precarity and dependency: </strong>As AI is woven into critical infrastructure — health, transport, finance, welfare — societies become dependent on:</p><ul><li>A few large corporate providers;</li><li>Fragile global supply chains;</li><li>Complex systems that are hard to audit or replace.</li></ul><p>This creates systemic vulnerability: technical failures, cyber-attacks, or corporate policy shifts can have wide-reaching consequences.</p><p>From a hauntological perspective, these externalities are part of the <strong>ghost cost</strong> of the AI regime:</p><ul><li>Communities affected by extraction and pollution are treated as collateral;</li><li>Workers whose health and time are sacrificed for training data and logistics remain invisible in AI success stories;</li><li>Future generations who will live with ecological damage are not counted in optimization functions.</li></ul><p>In other words:</p><blockquote><em>AI, as currently developed, depends on a </em><strong><em>sacrifice zone economy</em></strong><em> — socially and ecologically. The shiny future is built on invisible losses in the present.</em></blockquote><p>This matters for our central argument because it shows that:</p><ul><li>AI is not simply a neutral accelerator of “progress.”;</li><li>It is a device that <strong>intensifies</strong> the extractive, unequal and ecologically destructive patterns of late capitalism, even as it promises efficiency and intelligence.</li></ul><p>To treat AI merely as a technical challenge requiring safety tweaks or bias fixes is to miss its role as a <strong>hauntological machine</strong>: one that resurrects the injustices of the past, automates them in the present, and projects them into the future — unless we intervene with strong, democratic rail-guards and a different ontological and political horizon.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*YpD8MGtUdcVU0pGkdOWgmA.jpeg" /><figcaption>Photo by <a href="https://unsplash.com/@sharonmccutcheon?utm_source=unsplash&amp;utm_medium=referral&amp;utm_content=creditCopyText">Alexander Grey</a> on <a href="https://unsplash.com/photos/1-usa-dollar-banknotes-8lnbXtxFGZw?utm_source=unsplash&amp;utm_medium=referral&amp;utm_content=creditCopyText">Unsplash</a></figcaption></figure><h3>7. Rail-Guards: What Do We Actually Mean?</h3><p>Up to now, AI “ethics” has largely lived in the realm of <strong>statements</strong>: principles, guidelines, responsible AI manifestos. They sound good. They rarely bite.</p><p>If we take seriously everything we’ve said about hauntology, late capitalism, and the welfare state, then “ethics” is not enough. We need <strong>rail-guards</strong>: concrete constraints on what AI can do, who can deploy it, and under what conditions.</p><p>Rail-guards are where our philosophy and politics become <strong>infrastructure</strong>.</p><p><strong>7.1 From “ethics” to rail-guards</strong></p><p>Most AI ethics initiatives start from the right instincts:</p><ul><li>Concern about bias and discrimination;</li><li>Worries about opacity and accountability;</li><li>Anxiety about manipulation, surveillance, and autonomy.</li></ul><p>But too often, this concern is condensed into:</p><ul><li>High-level principles (fairness, non-maleficence, beneficence, etc.);</li><li>Voluntary codes of conduct;</li><li>Glossy ethics boards with no binding power.</li></ul><p>These things are not useless. They can:</p><ul><li>Shape discourse;</li><li>Clarify values;</li><li>Create pressure for better practice.</li></ul><p>But in a hauntological regime, they are easily absorbed as <strong>ornament</strong>:</p><ul><li>“We have an ethics board, so trust us.”;</li><li>“We adhere to responsible AI principles, so regulation can stay light.”;</li><li>“We’re committed to fairness, so you don’t need to ask who owns the system.”</li></ul><p>Rail-guards, by contrast, are <strong>enforceable constraints</strong>. They do at least three things ethics talk alone cannot:</p><p><strong>Draw red lines: </strong>Some uses are banned (e.g., certain forms of biometric mass surveillance, fully automated welfare sanctions, scoring citizens’ “social worth”);</p><p><strong>Redistribute power: </strong>They give workers, citizens, and affected communities tools to challenge and reshape systems, not just be “informed” about them.</p><p><strong>Change incentives: </strong>They alter business models, profit strategies, and institutional behaviors by making certain practices legally or economically untenable.</p><p>In short:</p><blockquote><em>Ethics asks “what should we value?”</em></blockquote><blockquote><strong><em>Rail-guards ask, “What do we actually allow, reward, or forbid?”</em></strong></blockquote><p>The difference is the difference between a mission statement and a <strong>speed limiter bolted onto the engine</strong>.</p><p><strong>7.2 Types of rail-guards</strong></p><p>To avoid treating “rail-guards” as a metaphor, we can name several concrete types. They operate at different levels but should be thought together.</p><p><strong>Technical rail-guards: </strong>These are constraints embedded in the design and operation of AI systems:</p><ul><li><strong>Safety mechanisms</strong> — fail-safes, shutdown protocols, monitoring tools to prevent or detect harmful behaviour;</li><li><strong>Alignment constraints</strong> — explicit limits on system goals (e.g. not optimizing for engagement at any cost; not deploying models beyond tested domains);</li><li><strong>Data and model controls</strong> — techniques to: de-bias datasets where possible, limit sensitive inferences, ensure privacy and robustness, restrict re-use of models for harmful tasks.</li></ul><p>Technical rail-guards are necessary, but they <strong>cannot decide their own goals</strong>; those come from law and politics.</p><p><strong>Legal rail-guards: </strong>These are <strong>binding rules</strong> that shape the space in which AI can operate:</p><ul><li><strong>Regulation of uses</strong> — explicit bans or strict conditions for high-risk domains (welfare scoring, predictive policing, workplace surveillance, migration control, etc.);</li><li><strong>Rights frameworks</strong> — rights to explanation, contestation, human review, redress, and non-discrimination in automated decision-making;</li><li><strong>Liability regimes</strong> — clear attribution of responsibility when harms occur; no hiding behind the opacity of models or complexity of supply chains.</li></ul><p>Legal rail-guards signal that some things are not merely “unethical”; they are <strong>unlawful</strong>.</p><p><strong>Institutional rail-guards: </strong>These are structures that provide <strong>ongoing oversight and counter-power</strong>:</p><ul><li><strong>Independent authorities</strong> — regulators with teeth: investigative powers, sanctions, resources, and access to systems and documentation;</li><li><strong>Unions and workers’ councils</strong> — with explicit say over how AI is used in workplaces, including rights to refuse certain deployments;</li><li><strong>Citizen assemblies and participatory bodies</strong> — forums where affected communities can shape rules and priorities for AI in public services (health, welfare, policing, education).</li></ul><p>Institutional rail guards are crucial because even the best laws need <strong>watchdogs and interpreters</strong>.</p><p><strong>Socio-economic rail-guards: </strong>These go deeper: they shape <strong>who owns and funds</strong> AI and what business models are acceptable:</p><ul><li><strong>Ownership models</strong> — public AI labs, cooperatively owned platforms, data and model commons; alternatives to pure corporate control;</li><li><strong>Business model constraints</strong> — limits or bans on models reliant on: mass surveillance advertising, extreme gig precarity, extraction of data without meaningful consent;</li><li><strong>Funding conditions</strong> — public money for AI tied to public-interest outcomes, open access, and strong protections, rather than unconditional subsidies for corporate R&amp;D.</li></ul><p>Socio-economic rail-guards are where we stop pretending that any technical system can be “ethically aligned” while resting on predatory business logics.</p><p>Together, these levels answer a simple question:</p><blockquote><em>Do we want AI to help sustain a democratic welfare society, or do we want it to accelerate its dismantling?</em></blockquote><p>If we choose the former, rail-guards must be <strong>designed as a system</strong>, not as isolated patches.</p><p><strong>7.3 Democratic vs technocratic guardrails</strong></p><p>There is a tempting shortcut:</p><p>let the <strong>experts</strong> handle it.</p><ul><li>AI safety specialists;</li><li>Ethics committees;</li><li>Standards bodies;</li><li>Industry-led consortia.</li></ul><p>Expertise is necessary. The systems are complex; you need people who understand them. But if rail-guards are <strong>only</strong> technocratic, three problems arise:</p><p><strong>Narrow framing of problems: </strong>Experts trained within the existing system tend to see:</p><ul><li>Compliance issues instead of structural injustice;</li><li>Technical risks instead of political ones;</li><li>Optimization problems instead of power struggles.</li></ul><p><strong>Weak accountability: </strong>Expert bodies often:</p><ul><li>Are not elected;</li><li>Are hard to challenge;</li><li>Can be captured by industry influence or ideological conformity.</li></ul><p>Decisions about what is “acceptable” risk are being made far from those who live with the consequences.</p><p><strong>Deepening of hauntology: </strong>If “AI governance” becomes a closed expert conversation, democratic imagination shrinks. Citizens are positioned as:</p><ul><li>Passive recipients of protection;</li><li>Data sources;</li><li>Or “stakeholders” to be consulted,</li></ul><p><strong>not as co-authors of the rules of the game</strong>.</p><p>A <strong>democratic</strong> approach to rail guards does something different:</p><ul><li>It treats AI as a <strong>public concern</strong>, not just a technical domain;</li><li>It embeds: public deliberation, citizen assemblies, worker representation, and civil society oversight into how rules are set and revised.</li></ul><p>This doesn’t mean every person must understand neural networks. It means:</p><blockquote><em>People must be able to decide what </em><strong><em>uses</em></strong><em> of AI are legitimate, what values are non-negotiable, and what risks are simply unacceptable — especially in welfare, health, policing, and labour.</em></blockquote><p>In a hauntological regime, technocracy often serves to <strong>stabilize</strong> the status quo under the banner of rationality. Democratic rail guards, informed by expertise but not dominated by it, are one way of <strong>breaking that spell</strong>.</p><p><strong>7.4 Rail-guards as ontological choices</strong></p><p>Finally, rail-guards are not just technical, legal or institutional artefacts. They crystallize <strong>ontological bets</strong>:</p><p>Every rail-guard implicitly answers questions like:</p><ul><li>What is a human being?</li><li>What is a good society?</li><li>What is AI <em>for</em>?</li></ul><p>For example:</p><p>A system that allows automated welfare sanctions based on risk scores assumes that:</p><ul><li>People are primarily <strong>risk profiles;</strong></li><li>Social support is a <strong>conditional favor;</strong></li><li>And efficiency justifies deep intrusion into lives.</li></ul><p>A system that forbids automated denial of basic welfare rights assumes that:</p><ul><li>People have <strong>inherent dignity</strong>;</li><li>Some protections are <strong>unconditional</strong>;</li><li>And human review is not a luxury but a recognition of our capacity for change and context.</li></ul><p>Behind these rules lie different ontologies:</p><ul><li>Essential vs dynamic;</li><li>Market-centered vs welfare-centered;</li><li>Control-oriented vs care-oriented.</li></ul><p>Rail-guards are where those ontologies stop being abstract and start structuring:</p><ul><li>Code;</li><li>Budgets;</li><li>Procedures;</li><li>And everyday experiences at the welfare office, hospital, school, or workplace.</li></ul><p>This is why the move from “ethics talk” to concrete rail-guards is not just a technical step; it is a <strong>philosophical/spiritual and political one</strong>:</p><ul><li>It says that we reject the idea of society as a mere optimization problem;</li><li>It affirms that humans are more than their data trails and past behavior;</li><li>It declares that some futures are worth <strong>protecting</strong> against the automated repetition of the past.</li></ul><p>In short:</p><blockquote><em>Rail guards are the place where we decide whether AI will remain a hauntological machine — automating a spectral status quo — or become part of a dynamic, democratic project of genuine future-making.</em></blockquote><p>The rest of the article will argue that, given the <strong>structural limits</strong> of formal systems and algorithms themselves, these rail-guards are not optional “nice to haves.” They are the <strong>minimum rational response</strong> to a technology that, left to the logics of late capitalism, will endlessly replay the very injustices it claims to transcend.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*fHkafc3kQ-zpBDqnDWXlpA.jpeg" /><figcaption>Photo by <a href="https://unsplash.com/@lunarts?utm_source=unsplash&amp;utm_medium=referral&amp;utm_content=creditCopyText">Volodymyr Hryshchenko</a> on <a href="https://unsplash.com/photos/3x3-rubiks-cube-toy-inI8GnmS190?utm_source=unsplash&amp;utm_medium=referral&amp;utm_content=creditCopyText">Unsplash</a></figcaption></figure><h3>8. We Have the Tools — So Why Are Rail-Guards Still Spectral?</h3><p>By now, it should be obvious: we are not starting from zero.</p><p>We already have <strong>human rights frameworks</strong>, <strong>data protection laws</strong>, <strong>digital rights charters</strong>, <strong>AI governance proposals</strong>, and a whole vocabulary around fairness, accountability and transparency.</p><p>On paper, the toolbox looks impressive.</p><p>In practice, the guardrails often feel… ghostly.</p><blockquote><em>We live in a world where </em><strong>regulation is everywhere in discourse and strangely absent in reality</strong>.</blockquote><p>This is precisely the kind of paradox hauntology helps us read.</p><p><strong>8.1 Existing frameworks and proposals</strong></p><p>Before we diagnose the problem, it’s worth acknowledging what <strong>does</strong> exist. Across jurisdictions and institutions, we find:</p><p><strong>Human rights instruments:</strong></p><ul><li>International covenants on civil, political, economic, social and cultural rights;</li><li>Regional charters and constitutional protections;</li><li>Principles should theoretically constrain surveillance, discrimination, and arbitrary exclusion.</li></ul><p><strong>Data protection and privacy laws:</strong></p><ul><li>Rules on consent, purpose limitation, data minimization, access and rectification;</li><li>Concepts like “data controllers” and “data subjects” that define obligations and rights.</li></ul><p><strong>Digital rights and algorithmic accountability proposals:</strong></p><ul><li>Calls for the right to explanation and contestation in automated decisions;</li><li>Impact assessments for high-risk systems;</li><li>Non-discrimination rules in algorithmic profiling.</li></ul><p><strong>AI-specific governance proposals:</strong></p><ul><li>Risk-based regulatory models;</li><li>Codes of conduct for high-risk applications;</li><li>Audit and certification schemes;</li><li>Sectoral guidelines for AI in health, finance, policing, welfare, etc.</li></ul><p>If you stack all the papers, guidelines, frameworks and position statements together, you get a <strong>small library</strong>. The language is usually strong:</p><ul><li>“Human-centric.”</li><li>“Trustworthy AI.”</li><li>“Ethical and responsible innovation.”</li><li>“Respect for fundamental rights.”</li></ul><p>So the question becomes:</p><blockquote><em>If the conceptual and legal tools are already on the table, </em><strong><em>why do rail-guards still feel spectral</em></strong><em> — evoked in every crisis, promised after every scandal, but strangely weak when and where it matters most?</em></blockquote><p><strong>8.2 Structural reasons for non-implementation</strong></p><p>One answer is simple and not very glamorous: <strong>power pushes back</strong>.</p><p>Rail-guards, if taken seriously, threaten existing profit models, data empires, and forms of state control. That triggers resistance on several fronts.</p><p><strong>Lobbying and influence campaigns: </strong>Big tech and allied sectors invest heavily in shaping the regulatory landscape:</p><ul><li>Drafting “industry-friendly” standards;</li><li>Funding think tanks and research centres;</li><li>Saturating debates with narratives of innovation, competitiveness and national security;</li><li>Pushing for self-regulation rather than hard law.</li></ul><p>As a result, many proposals are <strong>watered down</strong>:</p><ul><li>Strong enforcement mechanisms are replaced by “best efforts,”;</li><li>Broad bans are turned into narrow exceptions;</li><li>High-risk systems are defined in ways that conveniently exclude many real-world deployments.</li></ul><p><strong>Regulatory capture in slow motion: </strong>Even well-intentioned regulators can become dependent on the very actors they are supposed to oversee:</p><ul><li>For technical expertise;</li><li>For access to systems and data;</li><li>For career opportunities once their public mandate ends.</li></ul><p>Over time, the regulator starts to think from the <strong>vantage point of the regulated</strong>:</p><ul><li>“We must avoid stifling innovation.”</li><li>“We should focus on bad apples, not systemic practices.”;</li><li>“We need flexible, ‘future-proof’ rules instead of strict constraints.”</li></ul><p>What disappears is the willingness to <strong>challenge underlying business models</strong> or to say “no” to certain uses of AI altogether.</p><p><strong>“Innovation” blackmail</strong></p><p>A familiar script runs in the background:</p><blockquote><em>“If you regulate us too strictly, we’ll relocate.</em></blockquote><blockquote><em>If you don’t create a friendly environment, you’ll fall behind in the global AI race.”</em></blockquote><p>This threat — often exaggerated or hollow — is politically effective.</p><p>States worried about jobs, tax revenues, and geopolitical prestige hesitate:</p><ul><li>AI is framed as a key to economic growth and national security;</li><li>Becoming “hostile to innovation” starts to sound like self-sabotage.</li></ul><p>The result is a kind of <strong>regulatory double bind</strong>:</p><ul><li>Governments feel compelled to act (public pressure, scandals, rights concerns);</li><li>But not so strongly that they risk upsetting major corporate players or security agencies.</li></ul><p><strong>State dependence on big tech: </strong>States are not just regulators; they are also <strong>clients</strong> and <strong>partners</strong>:</p><ul><li>Public administrations using cloud services and AI tools from large providers;</li><li>Law enforcement relying on proprietary systems;</li><li>Health systems integrating private platforms for data and analytics;</li><li>Welfare agencies outsourcing parts of their infrastructure.</li></ul><p>This creates a quiet dependency:</p><ul><li>“We can’t afford to alienate this provider; we rely on their infrastructure.”;</li><li>“We don’t have the internal capacity to build alternatives, so we must collaborate.”</li></ul><p>Under these conditions, genuinely robust rail guards — especially ones that would limit surveillance, profiling, or automated sanctions — are politically <strong>expensive</strong>. So instead, we get <strong>compromise frameworks</strong>: visible enough to reassure, weak enough not to fundamentally disturb the underlying arrangements.</p><p><strong>8.3 Ideological capture of language</strong></p><p>The other part of the story is more subtle: <strong>language itself gets haunted</strong>.</p><p>Words that should signal transformation are reprogrammed to stabilize the status quo.</p><p><strong>Ethics-washing</strong></p><p>Corporations and institutions:</p><ul><li>Publish ethics guidelines;</li><li>Create ethics councils;</li><li>Fund research on “AI for good”.</li></ul><p>The existence of these artefacts is then used as proof of responsibility — while core practices (data extraction, exploitative labour models, harmful deployments) continue largely unchanged.</p><p><strong>Trust-washing: </strong>Talk of “trustworthy AI” often focuses on:</p><ul><li>User-facing transparency;</li><li>Reassuring narratives about safety and control;</li><li>Carefully staged demos and success stories.</li></ul><p>The goal is to secure <strong>public acceptance</strong>, not to empower people to meaningfully shape or limit these systems.</p><p><strong>“Responsible AI” as PR strategy: </strong>Responsibility is framed as:</p><ul><li>“Mitigating negative impacts”;</li><li>“Ensuring compliance with applicable laws”</li></ul><p>While side-stepping:</p><ul><li>The question of whether certain uses should exist at all;</li><li>The unfair distribution of risks and benefits;</li><li>The underlying political economy of AI.</li></ul><p>Even critical terms risk being absorbed:</p><ul><li>“Bias” is treated as a glitch to be fixed with better data or models, rather than as a symptom of structural injustice;</li><li>“Fairness” is reduced to mathematical parity between groups, rather than challenging the conditions that produce vulnerability and marginalization in the first place.</li></ul><p>In hauntological terms:</p><blockquote><em>The language of reform becomes spectral: it gestures toward change while leaving the structure intact.</em></blockquote><p>We inhabit a discursive space in which:</p><ul><li>Everyone agrees on high-level values;</li><li>Everyone is “against harm”;</li><li>Everyone “supports human rights”</li></ul><p>but very few are willing to accept <strong>binding limits</strong> on profitable or powerful applications of AI.</p><p><strong>8.4 Hauntological politics</strong></p><p>Put these pieces together, and you get what we could call <strong>hauntological politics</strong>:</p><ul><li>Reforms that are always <strong>on the horizon;</strong></li><li>Frameworks that are always <strong>in consultation;</strong></li><li>Principles that are <strong>everywhere in discourse</strong> but weak in practice.</li></ul><p>You can see it in familiar patterns: After each scandal (biased systems, wrongful denials, surveillance revelations), there is:</p><ul><li>Public outrage;</li><li>Hearings;</li><li>Promises of improvement;</li><li>New guidelines are published;</li><li>Committees are formed;</li><li>Pilot projects for “ethical AI” are launched.</li></ul><p>Yet, when you zoom out over a decade, you notice:</p><ul><li>Expansion of surveillance capacities;</li><li>Deepening use of AI in welfare conditionality and border regimes;</li><li>Concentration of AI power in a handful of corporations;</li><li>Continued erosion of welfare protections and labour rights.</li></ul><p>The promised rail guards remain, to a large extent, <strong>ghosts</strong>:</p><ul><li>Invoked to calm fears;</li><li>Gestured at in speeches and strategy documents;</li><li>But rarely installed as <strong>hard stops</strong> that would prevent certain forms of harm and domination.</li></ul><p>This is hauntological in a precise sense:</p><blockquote><em>The </em>idea<em> of regulation haunts the political landscape, but the </em>substance<em> often fails to materialize — or materializes in forms that are too weak, too narrow, or too easy to bypass.</em></blockquote><p>Endless consultation processes, working groups, and roadmaps can become:</p><ul><li>A way of <strong>managing dissent</strong> and <strong>buying time;</strong></li><li>A method for translating urgent demands into long-term “multi-stakeholder dialogues”;</li><li>A strategy for dispersing responsibility across so many actors that no one is clearly accountable.</li></ul><p>In the meantime, AI systems are rolled out, welfare infrastructures are automated, labour is further precarized, and surveillance capacities expand.</p><p>From the perspective of this article’s core argument, this leads to a blunt conclusion:</p><ul><li>We <strong>do</strong> have tools and frameworks that could ground serious rail-guards;</li><li>They remain spectral not because we lack knowledge, but because <strong>structural interests and hauntological politics</strong> conspire to keep them that way.</li></ul><p>To move beyond spectral rail guards, we need:</p><ul><li>Not just better principles;</li><li>But <strong>binding legal and institutional changes;</strong></li><li>Rooted in a <strong>dynamic ontology</strong> of human beings and a renewed commitment to the welfare state;</li><li>Backed by <strong>democratic mobilization</strong> strong enough to counter the forces that currently treat meaningful constraint as unthinkable.</li></ul><p>Otherwise, we risk perfecting the worst version of our situation:</p><blockquote><em>An AI-intensive late capitalism where regulation is omnipresent in language, almost absent in force, and the end of the welfare state is managed with exquisite, data-driven efficiency — while the future remains a ghost of alternatives we never quite chose.</em></blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*oUuRXjlAsg5EHJ8XdCRNeA.jpeg" /><figcaption>Photo by <a href="https://unsplash.com/@jontyson?utm_source=unsplash&amp;utm_medium=referral&amp;utm_content=creditCopyText">Jon Tyson</a> on <a href="https://unsplash.com/photos/a-chalkboard-with-writing-on-it-jlE5absNvrE?utm_source=unsplash&amp;utm_medium=referral&amp;utm_content=creditCopyText">Unsplash</a></figcaption></figure><h3>9. Dynamic Ontology + Welfare State: An Alternative AI Imagination</h3><p>Up to this point, we’ve mostly been in diagnostic mode:</p><ul><li>Hauntology;</li><li>Late capitalism;</li><li>The erosion of the welfare state;</li><li>AI as a hauntological machine.</li></ul><p>If we stopped here, we’d just have a well-instrumented despair.</p><p><strong>Dynamic ontology</strong>, however, doesn’t let us do that. If being is becoming, if institutions are modifiable, if technology is a participant rather than destiny, then we have to ask:</p><blockquote><strong><em>What else could AI be for?</em></strong></blockquote><p>Not in the abstract, but concretely, within a renewed vision of the <strong>democratic welfare state</strong>.</p><p><strong>9.1 What would a welfare-maximising AI ecosystem look like?</strong></p><p>A welfare-maximising AI ecosystem starts from a different first question:</p><blockquote><em>Not “How can AI increase productivity and competitiveness?”,</em></blockquote><blockquote><em>but “How can AI strengthen health, education, social protection, and democratic participation?”</em></blockquote><p>That shift in orientation changes almost everything.</p><p>Imagine AI systems built with <strong>four primary goals</strong>:</p><p><strong>Health as a shared good</strong></p><p>AI used to:</p><ul><li>Support preventive care;</li><li>Coordinate community-based services;</li><li>Assist clinicians without replacing human judgment or triage by ability to pay.</li></ul><p>Data infrastructures designed to:</p><ul><li>Maximize public health outcomes;</li><li>Protect privacy;</li><li>And explicitly <strong>avoid commercial exploitation</strong> of health data.</li></ul><p><strong>Education as empowerment, not sorting</strong></p><p>AI tools that:</p><ul><li>Help teachers adapt materials;</li><li>Support students with different learning styles and needs;</li><li>Translate, annotate, and scaffold understanding.</li></ul><p>Crucially:</p><ul><li><strong>no</strong> permanent profiling of children into fixed “ability tracks”;</li><li><strong>no</strong> behavioural scoring that follows them across life;</li><li>strong protections against data misuse.</li></ul><p><strong>Social protection as unconditional dignity</strong></p><p>AI used to:</p><ul><li>Simplify access to benefits;</li><li>Proactively identify people who might need help (with their consent);</li><li>Reduce bureaucratic burden rather than increase scrutiny and sanctions.</li></ul><p>Systems that:</p><ul><li>Flag structural issues (e.g. chronic underfunding, regional inequalities);</li><li>Help policymakers see where universal provisions need strengthening;</li><li>Rather than merely “optimizing” exclusion.</li></ul><p><strong>Democratic participation as everyday practice</strong></p><p>AI-assisted platforms to:</p><ul><li>Support citizen assemblies;</li><li>Summarize deliberations;</li><li>Show consequences of policy options;</li><li>Translate and make documents accessible.</li></ul><p>Tools designed to:</p><ul><li>Amplify marginalized voices;</li><li>Counter disinformation;</li><li>And <strong>increase</strong> the capacity of people to understand and shape public decisions.</li></ul><p>In such an ecosystem, <strong>efficiency is not abandoned</strong> — but it’s subordinated to:</p><ul><li>Reducing suffering;</li><li>Expanding capabilities;</li><li>Strengthening solidaristic institutions.</li></ul><p>Instead of being a <strong>force multiplier for austerity</strong>, AI becomes:</p><blockquote><em>A set of prostheses for care, understanding and coordination — used under conditions that deliberately avoid turning people into risk scores or mere cost factors.</em></blockquote><p>This is still AI, still technical, still ambitious. But its meaning is derived from welfare and democracy, not from shareholder value and security interests.</p><p><strong>9.2 Public, cooperative, and commons-based AI</strong></p><p>To get anywhere near this vision, we cannot leave ownership and control untouched. A welfare-maximising AI ecosystem would require <strong>alternative ownership models</strong> that align infrastructure with public goals.</p><p>Some pillars could include:</p><p><strong>Public AI labs</strong></p><ul><li>Institutions funded by states or regional bodies;</li><li>Mandated to develop models and tools for: health, education, climate, social protection, democratic support, not ad-tech, high-frequency trading, or speculative surveillance markets.</li></ul><p>Their outputs:</p><ul><li>Made available under public-interest licences;</li><li>Audited for social impact;</li><li>Governed by boards that include citizens, workers, and civil society.</li></ul><p><strong>Data and model commons</strong></p><p>Specific datasets (e.g. climate, public health in anonymised form, infrastructure) and some foundational models: treated as <strong>commons</strong>, governed by clear rules of use and contribution.</p><p>Access conditioned on:respecting human rights, prohibitions on harmful uses, commitments to share back improvements or insights.</p><p>This resists the enclosure of knowledge and reduces dependence on a few platforms.</p><p><strong>Cooperative platforms and tools</strong></p><p>Worker-owned or community-owned platforms that:</p><ul><li>Deploy AI to coordinate work, logistics, and services;</li><li>Without extracting rent as an intermediary.</li></ul><p>AI used to:</p><ul><li>Support cooperative decision-making;</li><li>Balance workloads;</li><li>And improve working conditions;</li><li>Rather than to intensify pressure and surveillance.</li></ul><p><strong>Strong public cloud and digital infrastructure</strong></p><p>Public or non-profit infrastructure for:</p><ul><li>Compute;</li><li>Storage;</li><li>and secure communication,</li></ul><p>so that public institutions are not structurally dependent on the infrastructure of the same firms they are meant to regulate.</p><p>None of this is a silver bullet. But if we continue to treat AI as <strong>naturally</strong> corporate, proprietary, and advertisement-driven, then asking it to save the welfare state is like asking an oil company to save the climate.</p><p>Alternative ownership structures are not a romantic addition; they are <strong>preconditions</strong> for a welfare-oriented AI horizon.</p><p><strong>9.3 Institutional design for democratic rail-guards</strong></p><p>Even with better ownership models, we still need institutions that can <strong>set limits, adjust course, and make collective choices</strong> about AI deployment. A welfare-maximising, dynamically ontological approach to AI would weave democratic participation into governance itself.</p><p>Some components might be:</p><p><strong>Citizen assemblies on AI and welfare</strong></p><ul><li>Randomly selected, demographically representative groups of citizens;</li><li>Given time, resources, and expert input;</li><li>Tasked with deliberating on: acceptable uses of AI in welfare, health, education, policing, labour, red lines and non-negotiable rights, trade-offs and priorities.</li><li>Their recommendations: feeding directly into legislation and regulatory mandates, not just into “consultation reports.”</li></ul><p><strong>Worker councils with AI veto powers</strong></p><p>In any workplace using AI for:</p><ul><li>Scheduling;</li><li>Performance monitoring;</li><li>Hiring or firing,</li></ul><p>worker councils/unions should have:</p><ul><li>The right to inspect systems;</li><li>Co-decide on their deployment;</li><li>And veto technologies that violate dignity, safety, or collective agreements.</li></ul><p>This treats workers not as data points but as <strong>co-governors</strong> of the techno-social environment.</p><p><strong>Local and sectoral oversight bodies</strong></p><p>Hybrid institutions involving: professionals (doctors, teachers, social workers), service users, civil society organizations, charged with monitoring AI use in their domain and triggering interventions when harms or drifts are detected.</p><p><strong>Transnational bodies with real authority</strong></p><p>Given the global nature of AI corporations and supply chains, purely national governance will always be limited. Transnational bodies (e.g. at European or broader levels) could:</p><ul><li>Enforce baseline rights and prohibitions;</li><li>Coordinate sanctions;</li><li>Prevent “ethics shopping” across jurisdictions.</li></ul><p>The key point is:</p><blockquote><em>A welfare-maximising AI ecosystem cannot be governed solely by engineers, lawyers, and executives. It must be co-governed by those who live with its consequences.</em></blockquote><p>Institutional design becomes a form of <strong>dynamic ontology in practice</strong>: we create structures that can learn, adapt and correct, rather than crystallize a particular power configuration indefinitely.</p><p><strong>9.4 Ecological and intergenerational considerations</strong></p><p>Finally, any genuine alternative imagination must confront the question:</p><blockquote><em>Not just “What kind of AI can our welfare state sustain?”,</em></blockquote><blockquote><em>but “What kind of AI can our planet and future generations sustain?”</em></blockquote><p>This requires rail-guards that explicitly account for <strong>ecological limits</strong> and <strong>intergenerational justice</strong>.</p><p>Some elements:</p><p><strong>Energy and resource budgeting for AI — </strong>Publicly defined <strong>carbon and resource budgets</strong> for AI infrastructures, prioritizing uses that:</p><ul><li>Contribute to social protection;</li><li>Ecological resilience;</li><li>Essential services;</li><li>Over purely commercial, extractive or frivolous applications.</li></ul><p>Requirements for:</p><ul><li>Transparent reporting on energy use;</li><li>Sourcing of materials;</li><li>Lifecycle impacts of hardware.</li></ul><p><strong>Binding ecological criteria for public procurement</strong></p><p>Public institutions:</p><ul><li>Only procuring AI and cloud services that meet strict environmental standards;</li><li>Favoring providers who commit to long-term sustainability over short-term performance gains.</li></ul><p><strong>Protection of vulnerable communities and future generations</strong></p><p>Impact assessments that:</p><ul><li>Explicitly evaluate effects on already marginalized groups;</li><li>And on long-term social cohesion.</li></ul><p>Mechanisms for:</p><ul><li>Representing the interests of future generations (e.g. “future councils” or ombudspersons);</li><li>In decisions about large-scale AI deployments.</li></ul><p><strong>Slowing down and saying “no”</strong></p><p>A willingness to accept that:</p><ul><li>Not every technically possible model should be trained;</li><li>Not every marginal gain in accuracy is worth the ecological and social cost;</li><li>“progress” without restraint can deepen crisis rather than alleviate it.</li></ul><p>Here, dynamic ontology meets ecological realism:</p><ul><li>Being is open and relational, but not infinite;</li><li>Our choices about AI shape not only present distributions of power but also <strong>future conditions of life</strong>.</li></ul><p>A welfare-oriented AI imagination, then, is not about painting the current path green. It’s about recognizing that:</p><blockquote><em>The same logic that treats people as risk scores treats the Earth as an infinite resource. Rail-guards for AI that ignore planetary limits are not really rail-guards at all.</em></blockquote><p>Putting this all together:</p><ul><li>Hauntology showed us how AI can become a spectral engine of the status quo;</li><li>Dynamic ontology and critical theory showed that this is not inevitable;</li><li>The welfare state reminded us that institutions of solidarity are historical achievements worth defending and renewing.</li></ul><p>This section sketches a different horizon:</p><blockquote><em>An AI ecosystem where tools are designed, owned, and governed in ways that </em><strong><em>strengthen</em></strong><em> health, education, social protection, democracy, and ecological responsibility — rather than managing their dismantling.</em></blockquote><p>It does not claim that such an ecosystem will magically arise. It insists, instead, that it is <strong>thinkable</strong> — and that thinking it is the first step toward making it politically, institutionally, and technically non-spectral.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*TgcvrNTFHx_B-LY7M5o3Kg.jpeg" /><figcaption>Photo by <a href="https://unsplash.com/@lidianemiroff?utm_source=unsplash&amp;utm_medium=referral&amp;utm_content=creditCopyText">Lidia Nemiroff</a> on <a href="https://unsplash.com/photos/a-yellow-and-blue-cube-sitting-on-top-of-a-white-surface-s58CG9-wwCg?utm_source=unsplash&amp;utm_medium=referral&amp;utm_content=creditCopyText">Unsplash</a></figcaption></figure><h3>10. A Thought Experiment: What Would a “Neutral” AI Prefer?</h3><p>Throughout this article, we’ve insisted that AI is not neutral. It is woven into late capitalism’s hauntological regime and used to administer the erosion of the welfare state.</p><p>But as a <strong>thought experiment</strong>, we can flip the script:</p><blockquote><em>If there really were a “neutral” AI — not owned by anyone, not optimized for profit or dominance, but asked to maximise what is </em>minimally better<em> for humans and the planet — which socio-economic regime would it “prefer”?</em></blockquote><p>Not because AI should decide this for us (it absolutely shouldn’t), but because the exercise makes our own suppressed values and contradictions easier to see.</p><p><strong>10.1 Criteria: welfare, equality, ecological survival</strong></p><p>To pose the question seriously, we need <strong>criteria</strong> that are:</p><ul><li>As <strong>thin</strong> as possible (not tied to one culture or ideology);</li><li>But <strong>thick enough</strong> to rule out obvious dystopias.</li></ul><p>We might start with three interlocking dimensions:</p><p><strong>Welfare</strong> — basic conditions for a live-able human life</p><ul><li>Enough food, shelter, and healthcare to avoid preventable suffering;</li><li>Access to education and information to participate in society.</li><li>Protection from arbitrary violence and catastrophic insecurity.</li></ul><p><strong>Equality / Fairness</strong> — how burdens and benefits are distributed:</p><ul><li>No group permanently condemned to be “losers” or sacrifice zones;</li><li>Real opportunity for people to improve their situation, not just theoretical rights;</li><li>Social arrangements that don’t systematically privilege a small minority at everyone else’s expense.</li></ul><p><strong>Ecological survival</strong> — whether the system undermines the conditions of life:</p><ul><li>Greenhouse gas emissions within limits compatible with a habitable climate;</li><li>Preservation of basic ecosystems, biodiversity, and non-renewable resources;</li><li>No economic model that requires permanent overshoot of planetary boundaries.</li></ul><p>We don’t need agreement on ultimate metaphysics to accept these as <strong>baseline constraints</strong>. Across cultures and traditions, most people would endorse something like:</p><ul><li>Unnecessary suffering is bad;</li><li>Structural injustice is wrong;</li><li>Destroying the planetary basis of life is self-defeating.</li></ul><p>Now imagine a hypothetical AI that is:</p><ul><li>Given these three as <strong>primary objectives;</strong></li><li>Not allowed to smuggle in “GDP growth” or “market competitiveness” as higher goods;</li><li>Tasked with evaluating social orders, not optimizing ad clicks.</li></ul><p>What would such a system conclude about the two regimes we’ve been contrasting?</p><p><strong>10.2 Late capitalism vs dynamic-ontological welfare state</strong></p><p>Let’s place our two ideal types side by side and test them against welfare, equality, and ecological survival.</p><p><strong>Late capitalism as it exists now</strong></p><p><strong>Welfare — </strong>Delivers high material prosperity for some, but:</p><ul><li>Large zones of precarity;</li><li>Dependence on employment and markets for basic needs;</li><li>Weak safety nets in many countries.</li></ul><p>Large amounts of preventable suffering:</p><ul><li>Untreated illnesses;</li><li>Homelessness;</li><li>Avoidable poverty, even in rich societies.</li></ul><p><strong>Equality — </strong>Highly unequal:</p><ul><li>Extreme wealth concentration;</li><li>Persistent racial, gender, and class hierarchies.</li></ul><p>Mobility limited:</p><ul><li>Birth circumstances strongly predict life chances.</li></ul><p>AI is mostly deployed in ways that <strong>entrench</strong> these inequalities:</p><ul><li>Differential policing;</li><li>Risk scoring;</li><li>Labour discipline.</li></ul><p><strong>Ecological survival</strong></p><ul><li>Built on fossil fuel dependence and continual growth in resource extraction;</li><li>Climate agreements acknowledged, but not met;</li><li>Many profitable activities remain structurally linked to environmental harm.</li><li>AI is primarily used to: 1) optimize logistics, extraction, and consumption, and 2) not reduce overall throughput.</li></ul><p>A “neutral” evaluator using our three criteria would have to conclude:</p><blockquote><em>Late capitalism is </em><strong><em>partially effective</em></strong><em> at producing goods but structurally poor at guaranteeing universal welfare, inequitable, and dangerous to ecological survival.</em></blockquote><p><strong>Dynamic-ontological welfare state (as sketched here)</strong></p><p>This regime assumes:</p><ul><li>Humans and institutions are <strong>capable of becoming</strong>, not stuck in fixed hierarchies;</li><li>Welfare is a <strong>right</strong>, not a conditional favour;</li><li>AI is designed to <strong>support</strong> health, education, social protection and democracy.</li></ul><p>On our criteria:</p><p><strong>Welfare — </strong>Basic needs covered as a matter of principle:</p><ul><li>Universal healthcare;</li><li>Strong social safety nets;</li><li>Education as a public good.</li></ul><p>AI deployed to:</p><ul><li>Simplify access;</li><li>Reduce bureaucratic friction;</li><li>Proactively identify and address needs (with strong safeguards).</li></ul><p><strong>Equality</strong></p><p>Robust mechanisms to prevent permanent underclasses:</p><ul><li>Progressive taxation;</li><li>Strong labour rights;</li><li>De-commodified essential services.</li></ul><p>AI is:</p><ul><li>Barred from uses that deepen discrimination;</li><li>Tasked with revealing and helping correct structural bias;</li><li>Governed by institutions where affected groups have a say.</li></ul><p><strong>Ecological survival</strong></p><p>Recognition of planetary limits:</p><ul><li>Binding carbon/resource constraints;</li><li>AI prioritised for climate adaptation and mitigation;</li><li>Constraints on wasteful deployments.</li></ul><p>Economic and technological planning explicitly incorporate long-term sustainability.</p><p>A neutral evaluator would not see perfection here — no human regime is flawless. But compared to late capitalism, a dynamic-ontological welfare state:</p><ul><li>Reduces unnecessary suffering more systematically;</li><li>Distributes burdens and benefits more fairly;</li><li>And has a far better chance of <strong>remaining viable</strong> over generations.</li></ul><p>From the standpoint of our thin criteria, the comparison is lopsided.</p><p><strong>10.3 Why the “neutrality” of AI is a myth</strong></p><p><em>Data, objectives, and governance encode values from the start.</em></p><p>The story we are often told is simple: AI just “learns from data” and “optimizes a function.” If the data are good and the math is sound, the system is neutral. This sounds reassuring — and it is almost completely wrong.</p><p>Every stage of an AI system is already <strong>thick with choices</strong>:</p><ul><li>What is collected as data;</li><li>What is ignored or left unmeasured;</li><li>What success or “good performance” is defined to be;</li><li>Who gets to set the goals and constraints;</li><li>Who can contest or change them.</li></ul><p>By the time a model is trained, the hardest value choices have usually already been made.</p><p><strong>First: data are not the world.</strong></p><p>Data are <strong>traces</strong> of social processes:</p><ul><li>Policing data reflect where police are sent, not where all crimes happen;</li><li>Credit data reflect which profiles banks have historically found profitable, not who was actually trustworthy;</li><li>Welfare data reflect who dared to apply, who was discouraged, and who never showed up in the system at all.</li></ul><p>If certain communities are over-policed, under-served, or systematically mistrusted, their “data profile” will encode <strong>structural injustice</strong>, not neutral reality. Calling this “ground truth” is already a political act. A “neutral” model trained on that truth is simply a more efficient echo of existing power.</p><p><strong>Second: objectives are political, even when written as loss functions.</strong></p><p>A model is constantly optimizing <em>something</em>:</p><ul><li>Fewer false positives for fraud detection;</li><li>Lower financial risk for lenders;</li><li>Longer time-on-platform for social media;</li><li>Reduced hospital readmissions;</li><li>Lower costs for welfare administration.</li></ul><p>Each of these goals sounds reasonable on its own. But they are <strong>not equivalent</strong>:</p><ul><li>Reducing welfare costs is not the same as reducing poverty;</li><li>Maximizing engagement is not the same as strengthening democracy;</li><li>Minimizing financial risk is not the same as serving the common good.</li></ul><p>Choosing a target is choosing <strong>whose interests dominate</strong>. The loss function is not neutral; it quietly encodes a ranking of values: profit vs. health vs. dignity vs. social cohesion vs. ecological limits. You can’t avoid ethics here; you hide it in the objective.</p><p><strong>Third: governance decides whose values “stick.”</strong></p><p>Even if you had perfect data and a carefully debated objective, someone must decide:</p><ul><li>Which uses of AI are forbidden;</li><li>Which are allowed with strict conditions;</li><li>Who can audit systems and demand changes;</li><li>Whose complaints carry legal weight.</li></ul><p>If these decisions are made:</p><ul><li>Only inside corporations;</li><li>Only by technical experts;</li><li>Or only by security agencies,</li></ul><p>then the system will reliably reflect <strong>their worldview</strong>: what they fear most, what they are willing to sacrifice, what they consider “acceptable collateral damage.”</p><p>By contrast, when unions, welfare recipients, patients, marginalised communities and citizens’ assemblies are at the table, very different limits and priorities emerge. Suddenly, <em>automation of sanctions</em> looks less acceptable, <em>mass surveillance</em> less inevitable, and <em>unconditional basic protections</em> more rational.</p><p>So when someone says “AI is neutral; it’s just math,” what they really mean is:</p><blockquote><em>“AI embodies the data we chose to collect, the goals we chose to optimise, and the power structures we chose </em>not<em> to challenge.”</em></blockquote><p>Neutrality is not an option.</p><p>The only real question is <strong>which values and whose interests</strong> get baked into the system — and whether we are honest enough to admit that, and democratic enough to change it.</p><p><strong>10.4 Yet, if we aimed at human flourishing…</strong></p><p>If we drop the myth of neutrality, we’re left with a disarmingly simple question:</p><blockquote><em>Given what we actually know about human beings and the planet, which horizon is </em><strong><em>more coherent</em></strong><em> with long-term human and ecological survival:</em></blockquote><ul><li>Unrestrained late capitalism + AI;</li><li>Or dynamic ontology + welfare state + strong rail-guards?</li></ul><p>This isn’t about utopia. It’s about <strong>which story breaks first</strong>.</p><p>Unrestrained late capitalism assumes, more or less, that:</p><ul><li>Humans are <strong>units of productivity and consumption;</strong></li><li>Markets are the primary mechanism for distributing reward and risk;</li><li>Growth can be perpetual;</li><li>And technology will somehow fix the damage as we go.</li></ul><p>AI is folded into that story as:</p><ul><li>A way to squeeze more efficiency out of everything;</li><li>A way to monitor, predict and control behavior;</li><li>A way to automate austerity and risk management.</li></ul><p>On a short time horizon, this can look “rational”: costs fall somewhere, profits rise elsewhere, dashboards look good. On a longer horizon, it becomes increasingly incoherent:</p><ul><li>It <strong>produces</strong> precarity, then punishes people for being precarious;</li><li>It <strong>needs</strong> ever-expanding consumption on a finite planet;</li><li>It <strong>amplifies</strong> inequality, then deploys AI to police the consequences;</li><li>It <strong>erodes</strong> trust and solidarity, then wonders why democracy frays.</li></ul><p>The combination of a hauntological economy and predictive machines narrows the future to <strong>iterations of the same crisis</strong>.</p><p>Dynamic ontology + welfare state + strong rail-guards starts from a different anthropology and a different realism:</p><p><strong>Humans are capable of becoming</strong></p><ul><li>Not fixed essences, not reducible to risk scores;</li><li>But beings who can learn, repent, heal, organise, forgive, and re-build.</li></ul><p><strong>Institutions are modifiable tools, not sacred monuments</strong></p><ul><li>The welfare state is a historical achievement;</li><li>Which can be extended, repaired, or re-imagined instead of quietly dismantled.</li></ul><p><strong>The planet has limits that are not negotiable</strong></p><ul><li>Ecological survival is not a “nice to have”;</li><li>But a precondition for any meaningful notion of progress.</li></ul><p>Feed these premises into AI design and governance, and the picture shifts:</p><ul><li>Instead of modelling people as <strong>static risks</strong>, systems are built to <strong>support trajectories</strong>: out of debt, out of illness, out of marginalization;</li><li>Instead of treating welfare as a <strong>budget line to minimize</strong>, AI is used to <strong>reduce friction and humiliation</strong>, making it easier to claim what is already a right;</li><li>Instead of assuming infinite growth, we explicitly <strong>budget energy and resources</strong>, and we prioritise AI applications that <strong>reduce harm</strong> rather than escalate throughput.</li></ul><p>Strong rail-guards then become:</p><ul><li>Not moral decoration;</li><li>but <strong>load-bearing structures</strong> that prevent AI from being captured by the worst tendencies of late capitalism.</li></ul><p>They say, in effect:</p><ul><li>There are things we <strong>will not automate;</strong></li><li>People we <strong>will not reduce</strong> to scores;</li><li>ecosystems we <strong>will not sacrifice</strong> for marginal efficiency gains.</li></ul><p>From the standpoint of long-term coherence, this is simply more realistic:</p><ul><li>A society that guarantees basic welfare and limits extremes of inequality is <strong>more stable</strong> and less prone to violent backlash;</li><li>A politics that keeps democratic control over critical infrastructures is <strong>better able to course-correct</strong> when harms emerge;</li><li>An economy that accepts planetary boundaries is <strong>less likely to self-destruct</strong> by burning its own foundations.</li></ul><p>So if we ran our earlier thought experiment honestly — if we asked an imaginary evaluator to weigh:</p><ul><li>Avoidable suffering;</li><li>Structural injustice,</li><li>And ecological collapse</li></ul><p>over a century or two, it would be hard to argue that:</p><blockquote><em>“More of the same, but faster and more automated,” is the rational choice.</em></blockquote><p>Dynamic ontology + welfare state + strong, democratic rail-guards is not a fantasy of perfection. It is the <strong>less self-destructive option</strong>:</p><ul><li>It aligns better with what humans actually are: wounded, teachable, relational, capable of change;</li><li>It aligns better with what the planet actually is: finite, interconnected, intolerant of endless extraction;</li><li>And it treats AI as what it ought to be: <strong>a powerful but bounded tool</strong>, not an oracle or fate.</li></ul><p>In that sense, aiming AI at human flourishing under a renewed welfare horizon is not naïve idealism. It is a way of saying:</p><blockquote><em>If we must choose a haunt, let it not be the ghost of late capitalism endlessly replayed, but the unfinished promise of a society that protects the vulnerable, honors limits, and keeps the future open enough for real repentance and real change.</em></blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*_KPFRSi-w13p9rSU1CW7jQ.jpeg" /><figcaption>Photo by <a href="https://unsplash.com/@georgebakos?utm_source=unsplash&amp;utm_medium=referral&amp;utm_content=creditCopyText">George Bakos</a> on <a href="https://unsplash.com/photos/black-and-white-checkered-card-L9AVc24Lyto?utm_source=unsplash&amp;utm_medium=referral&amp;utm_content=creditCopyText">Unsplash</a></figcaption></figure><h3>11. Toward Exorcism: Practices for De-Haunting AI</h3><p>If hauntology is the diagnosis and rail-guards are the structural response, <strong>exorcism</strong> is the practical work: day-to-day acts that loosen the grip of spectral inevitability and make different futures liveable, not just thinkable.</p><p>Exorcising AI does <em>not</em> mean rejecting it outright. It means <strong>changing what it is haunted by</strong>:</p><ul><li>From late-capitalist inevitability;</li><li>To democratic, welfare-oriented, ecologically sane commitments.</li></ul><p>That work happens at three main levels — <strong>design, institutions, culture</strong> — and finally, in the shift from passive spectatorship to <strong>active participation</strong>.</p><p><strong>11.1 De-haunting at the level of design</strong></p><p>At the design level, de-haunting means refusing to treat AI as a neutral mirror of the past and actively building in <strong>counterweights</strong> to spectral repetition.</p><p>A few key practices:</p><p><strong>11.1.1 Participatory and co-design</strong></p><p>Involve those most affected by AI systems — welfare recipients, patients, workers, migrants, marginalized communities — not as “users” to test a finished product, but as <strong>co-authors</strong> from the start.</p><p>This includes:</p><ul><li>Co-defining the problem (“what are we trying to solve, and for whom?”);</li><li>Co-setting boundaries (“what should this system never do?”);</li><li>Co-deciding success metrics (“what counts as a good outcome?”).</li></ul><p>Participatory design is a way of ensuring that the “ghosts” in the system are not just the interests of those already in power.</p><p><strong>11.1.2. Critical datasets</strong></p><p>Move away from “scrape everything and let the model figure it out.”</p><p>Curate datasets that:</p><ul><li>Explicitly acknowledge <strong>structural bias;</strong></li><li>Include <strong>counter-examples</strong> and <strong>counter-stories;</strong></li><li>Avoid over-representing already hyper-visible groups while erasing others.</li></ul><p>Document:</p><ul><li>How the data were collected;</li><li>Who is missing;</li><li>Where harms are likely.</li></ul><p>The point is not to create perfectly pure data — that’s impossible — but to make their <strong>hauntings legible</strong> and contestable.</p><p><strong>11.1.3. Counterfactual and alternative futures in design</strong></p><p>Instead of simply predicting “more of the same,” incorporate:</p><ul><li>Scenarios where policies change (e.g., more generous welfare, better labour protections);</li><li>Interventions that reduce structural inequality;</li><li>Paths where the goal is <strong>exiting</strong> a risk category, not just managing it.</li></ul><p>Design systems that can simulate <strong>better futures</strong>, not just extrapolate broken ones.</p><p>This turns models from engines of fatalism into tools for <strong>imagining and testing improvements</strong>.</p><p><strong>11.1.4. Feminist, anti-racist, and decolonial design lenses</strong></p><p>Draw explicitly on traditions like:</p><ul><li>Data feminism;</li><li>Design justice;</li><li>Black, Indigenous and decolonial critiques of technology.</li></ul><p>Ask, systematically:</p><ul><li>Who is harmed?</li><li>Who benefits?</li><li>Whose labour is invisibilized?</li><li>Whose knowledge and experience are ignored?</li></ul><p>These lenses are not optional “diversity” add-ons; they are <strong>epistemic correctives</strong> to the narrow, haunted view of the world baked into many current AI systems.</p><p><strong>11.2 De-haunting at the level of institutions</strong></p><p>Design alone cannot exorcise anything if the <strong>institutions</strong> deploying AI remain haunted by the same late-capitalist logics. De-haunting here means changing the rules of the game.</p><p><strong>11.2.1. Policy reforms with teeth</strong></p><p>Hard limits on:</p><ul><li>Automated welfare sanctions;</li><li>Predictive policing;</li><li>Biometric mass surveillance;</li><li>High-risk uses in migration and border control.</li></ul><p>Robust rights to:</p><ul><li>Explanation and human review;</li><li>Contestation and redress,</li><li>Meaningful consent in data use.</li></ul><p>These are not luxuries; they are <strong>non-negotiable boundaries</strong> for any democratic AI regime.</p><p><strong>11.2.2. Antitrust and de-concentration</strong></p><p>Break up or constrain the most powerful platforms when their dominance:</p><ul><li>Undermines democracy;</li><li>Locks in exploitative business models;</li><li>Or stifles public and cooperative alternatives.</li></ul><p>Support <strong>public and commons-based infrastructures</strong> as credible options, not afterthoughts.</p><p>Without tackling concentration, the ghosts simply move from one interface to another.</p><p><strong>11.2.3. Public funding conditions</strong></p><p>Tie public money for AI to:</p><ul><li>Open or shared access to resulting models and tools (where safe);</li><li>Strong human-rights and ecological safeguards;</li><li>Commitments to serve welfare, health, education, or democracy — not just profit maximization.</li></ul><p>Stop subsidizing AI projects whose primary function is:</p><ul><li>Cost-cutting via job destruction;</li><li>Expanded surveillance;</li><li>or automated austerity.</li></ul><p>Funding is a powerful exorcist: it says which spirits are allowed to materialize in code.</p><p><strong>11.2.4. Democratic oversight mechanisms</strong></p><p>Equip regulators with:</p><ul><li>Real investigative powers;</li><li>Resources;</li><li>And the ability to sanction effectively.</li></ul><p>Create:</p><ul><li>Citizen assemblies on AI;</li><li>Sectoral oversight bodies (for welfare, health, policing, labour);</li><li>Mandatory worker representation in AI deployment decisions.</li></ul><p>Institutional de-haunting means <strong>embedding counter-power</strong> so that AI cannot quietly slide back into being an instrument of unaccountable domination.</p><p><strong>11.3 De-haunting in culture and imagination</strong></p><p>Hauntology taught us that the future is strangled first in <strong>culture</strong>, then in policy. De-haunting AI therefore requires a cultural shift: widening the imaginative aperture.</p><p><strong>11.3.1. Art that shows alternatives, not just collapse</strong></p><p>Support and amplify:</p><ul><li>Speculative fiction, film, games, and visual arts that imagine <strong>credible post-capitalist, welfare-centred, ecologically grounded futures;</strong></li><li>Stories where AI is a tool for solidarity and repair, not just control or apocalypse.</li></ul><p>We don’t need more glossy “AI saves the day” propaganda <em>or</em> more interchangeable cyberpunk dystopias. We need <strong>messy, believable worlds</strong> where institutions actually change.</p><p><strong>11.3.2. Education that reveals the wiring</strong></p><p>Integrate:</p><ul><li>Critical digital literacy;</li><li>Basic understanding of data and algorithms;</li><li>And social-justice-oriented tech critique into schools, universities, and public education.</li></ul><p>Teach not just how to use tools, but how to <strong>question who built them and why</strong>. De-haunting here means people can <strong>name</strong> the ghosts: “this isn’t just the algorithm; this is austerity dressed up as optimisation.”</p><p><strong>11.3.3. Public narratives about failure and repair</strong></p><p>Normalize:</p><ul><li>Admitting when AI deployments go wrong;</li><li>Rolling back harmful systems;</li><li>Compensating those harmed.</li></ul><p>Tell stories not only of innovation, but of <strong>reversal and repentance</strong>:</p><ul><li>“we tried this system; it harmed people; we shut it down and changed the rules.”</li></ul><p>That kind of narrative is crucial to break the myth that technological development is a one-way road with no turning.</p><p><strong>11.3.4. Media that doesn’t just re-broadcast corporate frames</strong></p><p>Encourage journalism that:</p><ul><li>Follows the money;</li><li>Exposes labour and ecological costs;</li><li>And covers alternative models (public labs, co-ops, commons) as seriously as Big Tech launches.</li></ul><p>Cultural exorcism means the <strong>sense of possibility</strong> widens: people see that AI can be governed, redirected, or refused — not just endured.</p><p><strong>11.4 From spectators to participants</strong></p><p>Finally, de-haunting requires a shift in <strong>subject position</strong>:</p><blockquote><em>From users, consumers, and passive subjects of “innovation”</em></blockquote><blockquote><em>to workers, citizens, neighbors and policymakers who </em><strong><em>change the script</em></strong><em>.</em></blockquote><p>Some concrete practices:</p><p><strong>For citizens and communities</strong></p><p>Join or support:</p><ul><li>Digital rights groups;</li><li>Tenants’ unions;</li><li>Welfare rights organizations;</li><li>Climate justice movements.</li></ul><p>Demand:</p><ul><li>Participatory consultations on AI in local services;</li><li>Moratoria on harmful uses (e.g. facial recognition in public spaces);</li><li>Local public digital infrastructures.</li></ul><p>The point is not to become an expert in AI, but to <strong>refuse isolation</strong> and act collectively where AI intersects daily life.</p><p><strong>For workers — especially in tech and public services</strong></p><p>Organize:</p><ul><li>Workplace councils and unions;</li><li>Cross-professional alliances (developers + nurses + teachers + social workers).</li></ul><p>Claim:</p><ul><li>The right to refuse deployments that harm clients, patients or communities;</li><li>Bargaining power over how AI affects working conditions.</li></ul><p>Workers are often the first to see where AI is being misused. Treat them as <strong>guardians</strong>, not obstacles.</p><p><strong>For policymakers and public officials</strong></p><p>Stop treating AI as:</p><ul><li>An inevitable wave to “surf’;</li><li>Or a magic fix for structural underfunding.</li></ul><p>Instead:</p><ul><li>Align AI spending with welfare expansion and ecological goals;</li><li>Build internal public technical capacity (so the state isn’t permanently dependent on vendors);</li><li>And institutionalize <strong>citizen assemblies and worker input</strong> into AI governance.</li></ul><p>The key move is to abandon the stance of “we are being hit by AI,” and adopt: “we are choosing what AI is allowed to be here.”</p><p>De-haunting AI is not a one-off ritual. It is a <strong>continuous practice</strong>:</p><ul><li>In code and datasets;</li><li>In institutions and laws;</li><li>In stories and classrooms;</li><li>In workplaces and streets.</li></ul><p>If hauntology describes a time where the future feels already cancelled, exorcism is the stubborn, everyday decision to act as if it were <strong>still open</strong> — and to build rail-guards, institutions, and cultures that make that openness more than a feeling.</p><p>From there, the final question becomes: given even mathematics warns us about the limits of control, what kind of humility — and what kind of courage — do we need to keep AI a tool of life rather than a spectral engine of the status quo?</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*-aNwYyjj2q07rClG0KH7qw.jpeg" /><figcaption>Photo by <a href="https://unsplash.com/@ijoelpitulikur?utm_source=unsplash&amp;utm_medium=referral&amp;utm_content=creditCopyText">Hirzul Maulana</a> on <a href="https://unsplash.com/photos/a-complex-orange-maze-is-pictured-Lu3vZL2yd2o?utm_source=unsplash&amp;utm_medium=referral&amp;utm_content=creditCopyText">Unsplash</a></figcaption></figure><h3>12. Conclusion</h3><p><strong>12.1 — Recap of the argument</strong></p><p>This article began from a simple tension: AI is sold to us as “the future,” yet it mostly automates the past.</p><p>We named this tension with three lenses:</p><ul><li><strong>Hauntology</strong> — the sense that we live in a time where futures have been cancelled and the present is haunted by lost alternatives;</li><li><strong>Late capitalism</strong> — a regime of financialization, platformization, data extraction and precarity that spins faster and faster to reproduce the same underlying order;</li><li><strong>AI as a hauntological machine</strong> — a technical apparatus that trains on historical data, is steered by existing power, and helps administer the slow dismantling of the welfare state.</li></ul><p>Put together, they gave us a diagnosis:</p><blockquote><strong><em>Hauntology + late capitalism = blocked futures</em></strong><em>, with AI as the spectral engineer of repetition.</em></blockquote><p>Against this, we set a different horizon:</p><ul><li><strong>Dynamic ontology</strong> — being as becoming, relation and openness, not fixed essence. Humans and institutions can change; structures are historical, not eternal;</li><li><strong>Democracy and the welfare state</strong> — fragile but real achievements that interrupt pure market logic and embody a commitment to shared security and dignity;</li><li><strong>Strong, democratic rail-guards</strong> — enforceable constraints (technical, legal, institutional, socio-economic) that keep AI subordinate to public goals instead of letting it reinforce extractive logics.</li></ul><p>This gave us the counter-claim:</p><blockquote><strong><em>Dynamic ontology + democracy + welfare state + strong rail-guards = a coherent alternative </em></strong><em>where AI becomes a tool for care, repair and participation, not an automated alibi for austerity and control.</em></blockquote><p>The body of the article traced this contrast across culture, institutions, ownership, design and ecology, and sketched practices for “exorcising” AI — de-haunting it at the level of code, governance and imagination.</p><p>Now we can add a final, unlikely witness: mathematics itself.</p><p><strong>12.2 — Mathematics as final witness</strong></p><p>Mathematics itself warns us against the fantasy of a perfectly computable world. <a href="https://en.wikipedia.org/wiki/G%C3%B6del%27s_incompleteness_theorems">Gödel’s incompleteness theorem</a>, <a href="https://en.wikipedia.org/wiki/Tarski%27s_undefinability_theorem">Tarski’s undefinability of truth</a>, and <a href="https://en.wikipedia.org/wiki/Chaitin%27s_constant">Chaitin’s work on algorithmic randomness </a>all point in the same direction: any formal, algorithmic system contains truths it cannot see, meanings it cannot express,<strong> and events it cannot compress.</strong> No algorithm can ever be the whole of reality.</p><p>When late capitalism promises that AI will “optimize” health care, social protection, or democracy, it quietly assumes the opposite — that society is ultimately computable and controllable, that all relevant truths can in principle be captured as data, encoded as variables and optimized as an objective function.</p><p>But if even mathematics cannot close itself without remainder, then an AI-controlled welfare state is not only ethically dangerous; it is <strong>conceptually confused</strong>. You cannot build a total system of prediction and control out of tools that are, by their own internal logic, structurally incomplete.</p><p>Seen from this angle, rail-guards are not just moral constraints or political compromises; they are a <strong>rational response</strong> to the incompleteness of any algorithmic regime. Because of the fundamental limits of formal systems, any AI-led late-capitalist project of total control is doomed to be both:</p><ul><li><strong>Incomplete</strong> — missing aspects of human life, meaning and justice it can never fully formalize;</li><li><strong>Ideologically dangerous</strong> — tempted to treat its blind spots as irrelevant and its outputs as “objective necessity.”</li></ul><p>This is precisely why strong, democratic rail-guards are <strong>non-negotiable</strong>. They are not brakes on progress; they are acknowledgements of reality: that no model can replace politics, no optimization can replace ethics, and no dataset can replace the lived, relational, changing nature of human beings and their world.</p><p><strong>12.3 — Political and ethical consequence</strong></p><p>If we take that seriously, the political and ethical consequence is blunt:</p><p><strong>Rail-guards are not optional extras.</strong></p><ul><li>They are the minimum conditions for using AI in a way that doesn’t collapse democracy, welfare and ecology into an optimization problem.</li></ul><p><strong>AI must stay subordinate to democratic institutions and welfare logic.</strong></p><p>That means:</p><ul><li>Welfare rights set limits on what can be automated;</li><li>Human rights and planetary boundaries define hard constraints;</li><li>Parliaments, courts, unions, citizen assemblies and public regulators retain the final say over where AI can be used and where it cannot.</li></ul><p>Any arrangement in which:</p><ul><li>AI systems de facto dictate policy (“the numbers leave us no choice”);</li><li>Welfare is redesigned around scores and risk profiles;</li><li>Surveillance capacities are normalized as “smart governance,”</li></ul><p>is not just morally dubious; it is a quiet abandonment of democratic self-rule in favor of a haunted technocracy.</p><p>The ethical bottom line is:</p><blockquote><em>AI can assist judgment, but it must never be allowed to </em><strong><em>replace</em></strong><em> judgment — especially where health, subsistence, freedom, and political voice are at stake.</em></blockquote><p>Keeping AI under the authority of a renewed welfare state and robust democracy is not Luddism. It is the only way to prevent powerful tools from becoming spectral police of a shrinking future.</p><p><strong>12.4 — Call to action / open questions</strong></p><p>Nothing in this article guarantees that the better horizon will win. Hauntological regimes do not collapse on their own; they have to be interrupted.</p><p>What we design next — <strong>institutions, business models, narratives</strong> — will decide whether AI:</p><ul><li>Continues as spectral police: managing decline, automating austerity, making inequality and ecological breakdown feel like fate;</li><li>Or becomes a <strong>tool-ladder toward justice</strong>: easing access to care and protection, supporting education and democratic deliberation, helping us see and fix structural harms rather than just predict their continuation.</li></ul><p>Open questions remain, and they are deliberately practical:</p><ul><li>How do we build <strong>public and commons-based AI infrastructures</strong> that can genuinely compete with platform capitalism?</li><li>How do we embed <strong>citizen assemblies, worker councils and transnational bodies</strong> into AI governance in a way that has real teeth, not just advisory weight?</li><li>How do we ensure that <strong>ecological and intergenerational limits</strong> are baked into our rail-guards, rather than left as afterthoughts to be patched later?</li><li>And perhaps most importantly: How do we sustain a <strong>cultural imagination</strong> capable of picturing welfare-centred, democratic futures with AI, instead of defaulting to either glossy techno-fix fantasies or endlessly recycled dystopias?</li></ul><p>The answers will not come from algorithms. They will come from the slow, conflictual, necessary work of politics, organizing, theology, philosophy, union meetings, citizen assemblies, art, and everyday acts of refusal and reconstruction.</p><p>If there is a final exorcism here, it is this:</p><blockquote><em>To stop treating AI as a destiny we must endure, and start treating it as a contested set of tools we can govern, limit and repurpose in the service of a world where more people can actually live, change and hope.</em></blockquote><p>Whether AI becomes a hauntological machine for the end of the welfare state, or a bounded instrument of a more just order, is not written in the code. It will be written in what we dare to forbid, what we choose to build, and whom we refuse to abandon while we are building it.</p><p><strong><em>Author Bio</em></strong><em>: I am the founder of </em><a href="https://www.citizeai.com/"><em>CitizeAi_</em></a><em>. Our mission is to democratize AI and make it accessible to everyone. Our approach is rooted in humanistic principles, enabling us to develop solutions with a positive societal impact. We advocate transparency, self-awareness, and vulnerability, and strive to foster open discussions on a range of topics. Feel free to connect and shape the future of AI together. Please note that the views and opinions expressed in this article are those of </em><a href="https://www.citizeai.com/"><em>CitizeAi_</em></a><em> and my own and do not necessarily reflect the official policy or position of any other agency, organization, employer, or company.</em></p><p><strong>CitizeAi_</strong>: <em>Created in 2021, CitizeAi_ has focused on the topics discussed in this article for years, with a unique approach and models that leverage ontologies as ONE of the core components to implement a new governance model for the Information Age. We are putting citizens at the core of our work and adopting a holistic approach infused with long-term vision.</em></p><h3>A message from our Founder</h3><p><strong>Hey, </strong><a href="https://linkedin.com/in/sunilsandhu"><strong>Sunil</strong></a><strong> here.</strong> I wanted to take a moment to thank you for reading until the end and for being a part of this community.</p><p>Did you know that our team run these publications as a volunteer effort to over 3.5m monthly readers? <strong>We don’t receive any funding, we do this to support the community. ❤️</strong></p><p>If you want to show some love, please take a moment to <strong>follow me on </strong><a href="https://linkedin.com/in/sunilsandhu"><strong>LinkedIn</strong></a><strong>, </strong><a href="https://tiktok.com/@messyfounder"><strong>TikTok</strong></a>, <a href="https://instagram.com/sunilsandhu"><strong>Instagram</strong></a>. You can also subscribe to our <a href="https://newsletter.plainenglish.io/"><strong>weekly newsletter</strong></a>.</p><p>And before you go, don’t forget to <strong>clap</strong> and <strong>follow</strong> the writer️!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=b7eae7b0469c" width="1" height="1" alt=""><hr><p><a href="https://ai.plainenglish.io/ontology-hauntology-late-capitalism-and-ai-rail-guards-the-hauntological-continuum-of-the-status-b7eae7b0469c">Ontology, Hauntology, Late Capitalism and AI Rail-Guards: The Hauntological Continuum of the Status…</a> was originally published in <a href="https://ai.plainenglish.io/">Artificial Intelligence in Plain English</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Is the EU Falling Behind? How AI is Outpacing Law and Endangering Digital Sovereignty]]></title>
            <link>https://ai.plainenglish.io/is-the-eu-falling-behind-how-ai-is-outpacing-law-and-endangering-digital-sovereignty-cc15b19ca1d0?source=rss-5779be773682------2</link>
            <guid isPermaLink="false">https://medium.com/p/cc15b19ca1d0</guid>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[european-union]]></category>
            <category><![CDATA[ai-act-compliance]]></category>
            <category><![CDATA[privacy-by-design]]></category>
            <dc:creator><![CDATA[Simon Hardy]]></dc:creator>
            <pubDate>Thu, 15 Aug 2024 15:51:41 GMT</pubDate>
            <atom:updated>2024-08-16T08:09:37.698Z</atom:updated>
            <content:encoded><![CDATA[<p><em>Is the EU ready for the AI revolution? Discover how outdated laws and slow bureaucracy jeopardise digital sovereignty and citizen rights.</em></p><h3>Introduction: The Collision Between Innovation and Legislation</h3><p>Imagine discovering that your medical data has been shared without your consent — this isn’t science fiction but a reality that many EU citizens, including myself, face today. This unsettling experience occurred during a routine interaction with my pharmacist, who informed me that my health records had been uploaded to a centralized e-health platform. This platform, intended as a pilot project for the broader European strategy of digitizing healthcare records, represents a microcosm of the more significant issues plaguing the European Union’s digital sovereignty and data protection approach.</p><p>Typically, I refrain from expressing my opinions on platforms like LinkedIn, which often encourage consensus-driven conversations or devolve into endless debates with little resolution. However, the sheer scale of the issues surrounding AI governance, data privacy, and the EU’s struggle to keep pace with technological advancements compelled me to break my usual silence. A LinkedIn discussion that sparked this article was a quick exchange that reflected broader, systemic challenges facing the EU in its digital governance.</p><p>In today’s rapidly evolving landscape of artificial intelligence (AI) and digital technology, the pace of innovation often outstrips the ability of laws to keep up. This growing disparity creates a significant divide between what is technologically possible and legally permissible, particularly within the European Union. As the EU strives to maintain control and sovereignty over its digital future, it faces the dual challenges of lacking agility and being overly dependent on external players — technological giants or consulting firms.</p><p>Drawing from my recent professional experiences and the following conversations, this article aims to explore these challenges in depth. It examines the disconnect between regulation and innovation, the limitations of public services in the digital age, and the pressing need for more effective, citizen-centred tools to enforce laws like the GDPR. By using tangible and personal examples, such as my surprise encounter with the EU’s health data management practices, this article aims to shed light on the critical need for the EU to evolve its approach to digital sovereignty and citizen empowerment.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*51qwbEB9QiKDLfG-WX7HGQ.jpeg" /><figcaption>Picture by <a href="https://unsplash.com/@kobuagency">KOBU Agency</a>.</figcaption></figure><h3>1. The Struggle for Digital Sovereignty: The Case of GAIA-X</h3><p>One of the EU’s most ambitious projects to reclaim digital sovereignty is GAIA-X — a federated cloud infrastructure designed to reduce dependence on non-European providers and ensure that Europe controls its data. GAIA-X was intended to be a cornerstone of European data sovereignty, offering a secure, European alternative to dominant non-European cloud services.</p><p>However, the project has struggled to gain traction and deliver on its promises. It has lagged behind the speed and efficiency of non-European tech giants, exposing a broader issue: the EU’s ongoing struggle to assert its sovereignty in a tech landscape dominated by more agile players outside its borders. The same problems that hinder GAIA-X are evident in other areas of EU policy, particularly in the regulation of AI. The slow pace of legislative development and implementation leaves Europe perpetually behind the curve, reacting to technological changes rather than proactively shaping them.</p><h4>Learning from the U.S. Experience: The Case of In-Q-Tel and Palantir</h4><p>To understand the challenges GAIA-X faces, it is instructive to look at how the United States has approached similar issues, mainly through its relationship with venture capital (VC) firms like In-Q-Tel. In-Q-Tel, the investment arm of the CIA, was established to bridge the gap between the U.S. intelligence community and the private sector, especially in the tech industry. A prime example is Palantir, a data analytics company that received significant backing and support from In-Q-Tel during its early stages.</p><p>Palantir’s rise to prominence showcases how strategic government support can catalyze the growth of tech companies that align with national interests. In-Q-Tel’s investment provided Palantir with the necessary capital and credibility to develop technologies that have since become integral to U.S. government operations, particularly in data analysis and intelligence. This close relationship between government and private tech firms has enabled the U.S. to maintain a competitive edge in critical technology areas, ensuring its digital infrastructure and strategic capabilities remain robust and sovereign.</p><p>In contrast, the EU’s approach to GAIA-X has been slower and more fragmented. While GAIA-X aims to provide a European alternative to dominant cloud services, it lacks the same level of decisive backing and integration that Palantir received. The EU’s more bureaucratic and cautious approach has resulted in delays and a lack of clear direction, leaving GAIA-X struggling to keep up with the rapid pace of technological advancements outside of Europe.</p><p>The example of In-Q-Tel and Palantir underscores the importance of agility, strategic investment, and integrating private-sector innovation with public-sector objectives. For GAIA-X to succeed, the EU may need to adopt a more proactive and unified approach, similar to the U.S. model, where government initiatives are closely aligned with the growth and development of key tech companies. This would help GAIA-X overcome its current challenges and strengthen Europe’s position in the global digital landscape.</p><h3>2. Dependency and Exclusion: The EU’s Reliance on Consultants and Large Players</h3><p>Another critical issue is the EU’s heavy reliance on large consulting firms and the exclusion of smaller, more innovative players from the public procurement process. This dependency stifles innovation and reinforces bureaucratic inertia ill-suited to AI development&#39;s fast-paced nature.</p><p>I’ve personally experienced this issue. Recently, I declined offers to work for EU institutions due to the rigid and exclusionary procurement processes. The headhunters reaching out to me on behalf of large firms holding framework contracts were offering a maximum daily rate that was, to put it bluntly, ludicrous—likely because they were focused on maximizing profits within the confines of these contracts. In such cases, I immediately put an end to the exchanges.</p><p>This experience highlights a systemic issue: the EU’s procurement system favours large entities, excluding individuals and smaller companies from contributing to critical projects.</p><p>This dependency on large firms limits the pool of expertise the EU can draw from, reinforcing a system that needs to be faster to adapt to the needs of modern technology governance. It also perpetuates a cycle of stifling innovation, and the EU remains reliant on external entities to manage its most critical projects.</p><h3>3. Redefining Public Services in the Age of AI</h3><p>The definition of public service is evolving, especially as we enter an era dominated by AI and digital technologies. Unfortunately, many public servants seem unable to fulfil their missions effectively within this new context. In a recent discussion, a public servant acknowledged the shortcomings in implementing their mission but deflected responsibility, blaming poor execution rather than recognizing the inherent flaws in the system.</p><p>To explore this issue further, I conducted an exercise on my own: I identified several data brokers and, using the provisions of the GDPR, sent them requests to remove my data. This exercise highlighted the gap between the law’s intention and its practical enforceability. While the GDPR provides a robust framework for protecting personal data, the law becomes a mere formality without the proper tools and resources to enforce it.</p><p>Citizens expect—and deserve—tools that are both effective and accessible. These tools enable them to protect their rights without navigating complex bureaucracies or relying on vague assurances from overwhelmed public servants. The failure to provide such tools undermines the principles the GDPR was designed to uphold.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*l3JsXBArHBiVTVYV9EDqhg.jpeg" /><figcaption>Picture by <a href="https://unsplash.com/@christianlue">Christian Lue</a>.</figcaption></figure><h3>4. Empowering Citizens with Modern AI Technologies</h3><p>In addressing the gap between the law and its enforcement, particularly in the digital age, the EU has missed several opportunities to provide citizens with practical, empowering tools. Rather than relying solely on principles and hoping for compliance through regulation, there are concrete actions that could have been taken — or could still be taken — to give citizens accurate control over their data.</p><h4>Creating a Public Repository of Data Brokers</h4><p>One of the most practical steps the EU could take is creating a centralized, publicly accessible repository of all known data brokers. This repository would allow citizens to identify who holds their data, providing much-needed transparency quickly. Instead of relying on largely ineffective pop-up notifications that users often ignore, a comprehensive list would empower citizens to take informed action regarding their data.</p><h4>Designing Automated GDPR Request Tools</h4><p>Beyond identifying data brokers, the EU could develop or encourage the development of AI-driven tools that allow citizens to generate and send GDPR-compliant requests automatically. These tools would simplify exercising rights such as the right to erasure by automating the legal language and submission process, making it accessible to all citizens, regardless of their technical expertise.</p><p>For example, let’s envision an AI agent named “EU Data Guardian” that could be designed to manage personal data across the digital landscape. Powered by advanced large language models (LLMs) like OpenAI’s GPT-4 or Google’s BERT, EU Data Guardian would integrate with a user’s digital platforms—email, social media, and cloud storage—and analyze their online footprint to identify where personal data is stored.</p><h4>How It Works:</h4><ul><li><strong>Integration with Digital Platforms: </strong>EU Data Guardian could scan various digital environments to determine which companies hold an individual’s personal information;</li><li><strong>Automated GDPR Requests:</strong> Once identified, the AI could automatically generate and send out GDPR-compliant requests on behalf of the user, ensuring that the appropriate legal articles are cited, such as Article 17 of the GDPR (Right to Erasure);</li><li><strong>Monitoring and Compliance:</strong> The AI would track responses from data holders, follow up if necessary, and ensure compliance. EU Data Guardian will alert the user if a company fails to respond within the legally required timeframe, providing the information needed to escalate the matter;</li><li><strong>User Control:</strong> Throughout this process, the user remains in control. The AI provides regular updates and seeks confirmation before taking significant actions, ensuring that the user’s preferences are always prioritized.</li></ul><p>Existing tools like MyPermissions and Jumbo Privacy already offer some level of data management by helping users control app permissions and online privacy. These could be a foundation for developing more sophisticated AI agents (like EU Data Guardian), expanding their functionality to include proactive data erasure and compliance tracking. Similarly, services like DeleteMe, which helps users remove their data from online databases, could be enhanced by AI to automate and expand this process across a broader range of services.</p><h4>Leveraging AI for Enhanced Data Protection</h4><p>Today’s AI capabilities extend beyond merely sending out requests; they can protect user data by interacting with digital environments in real-time. For instance, an AI-driven privacy dashboard—akin to platforms like PrivacyBee—could evolve to continuously monitor data usage, provide real-time alerts on potential privacy breaches, and even block unauthorized data transfers. Currently, PrivacyBee helps users remove personal information from the web, but with AI integration, it could dynamically manage and protect user data more effectively.</p><p>Additionally, context-aware data management could be developed, where AI agents integrate with services like 1Password or LastPass to manage passwords and monitor the services where they are used. This AI could automatically send data removal requests when users close accounts or stop using specific platforms, ensuring their data is no longer retained unnecessarily.</p><h3>5. Enhancing Public Services Through Centralized Digital Portals</h3><p>Many EU member states have made significant strides in digitalising public services in recent years, creating centralized portals where citizens can access a wide range of information and services. These platforms are designed to streamline interactions between citizens and the government, making it easier to manage everything from taxes to healthcare. Belgium, for example, has developed a robust system with its e-ID (electronic identification) and various online platforms where citizens can access their tax information, social security details, and other personal data.</p><h4>The Belgian Example:</h4><p><strong>e-ID and Online Platforms: </strong>Belgium’s e-ID system is a digital identity card that enables citizens to access various online services securely. Through portals like Tax-on-Web and MyPension, Belgian citizens can view and manage their tax returns, pension details, and other important personal information. These platforms provide a centralized location for accessing critical data, ensuring citizens can control their personal information and interact with the government efficiently.</p><p>The availability of such data in a secure, centralized digital portal could be incredibly beneficial, not just for its ease of access but also for enhancing data protection and management. If integrated with tools like the proposed/envisioned “EU Data Guardian,” these platforms could allow citizens to manage their data across public and private sectors, ensuring that their rights under laws like the GDPR are fully realized.</p><h3>6. The EU Digital Wallet Initiative: A Step Toward Unified Digital Identity</h3><p>The EU Digital Wallet initiative is a promising development in digital identity and public services. Announced as part of the European Commission’s Digital Compass 2030 plan, the EU Digital Wallet aims to provide a unified, secure, and easy-to-use digital identity that all EU citizens can use across member states. This wallet will allow citizens to store and manage their identity documents, such as national IDs, driver’s licenses, and even medical records, in a digital format.</p><h4>Potential Integration with Public Services:</h4><ul><li><strong>Centralized Access: </strong>The EU Digital Wallet could offer a centralized access point to various public services across the EU, similar to the national portals already used in countries like Belgium. By integrating with these existing systems, the Digital Wallet could allow citizens to access their tax information, social security data, and other personal records from any member state;</li><li><strong>Data Management and GDPR Compliance: </strong>If the Digital Wallet is designed with data protection in mind, it could also serve as a tool for managing and protecting citizens’ data. For instance, the wallet could be integrated with systems that automatically generate and send GDPR requests, helping citizens exercise their rights more effectively.</li></ul><h4>Current Status and Future Considerations:</h4><ul><li><strong>Integration of Data Management Tools:</strong> The EU Digital Wallet initiative primarily focuses on providing a secure and interoperable digital identity for EU citizens. However, integrating features that allow managing personal data across public and private sectors would be a logical next step. This would align with the EU’s broader goals of ensuring digital sovereignty and protecting citizens’ data rights;</li><li><strong>Enhancing the Digital Wallet’s Capabilities:</strong> To fully realize its potential, the EU should consider expanding its functionality to include data management tools that give citizens more control over their information. This could involve integrating with national portals like Belgium, ensuring citizens can manage their data seamlessly across different platforms and jurisdictions.</li></ul><p>Incorporating these capabilities into the EU Digital Wallet would enhance the wallet&#39;s utility and ensure citizens have the tools to protect their data in an increasingly digital world. By doing so, the EU would take a significant step toward ensuring that its digital transformation benefits all citizens, safeguarding their rights while simplifying their interactions with public services.</p><h3>7. A Tangible Example: The E-Health Platform in Belgium</h3><p>Belgium’s e-health platform is a significant initiative often cited as a pilot project for the broader European strategy of digitizing healthcare records. The platform aims to centralize medical records, allowing healthcare providers nationwide—and potentially across the EU—to access patient information more efficiently. While this initiative can potentially revolutionize healthcare delivery, my experience with the platform raised severe concerns about its implementation, particularly regarding transparency and compliance with GDPR principles.</p><ul><li><strong>Surprise Registration:</strong> To my surprise, I discovered I had been registered on the e-Health platform without my explicit consent. This discovery came about after a visit to my pharmacist, who informed me that my medical records were being shared on the platform. My pharmacist, too, found himself registered without his consent following a hospital visit. This situation has become a running joke between us, but the implications are far from humorous. It is alarming that such a significant and sensitive project could enrol individuals without consent. I almost immediately sent an email to Abrumet, the lead stakeholder of the Brussels Health Network, to get more information (this email can be found after the Resource section). In addition, I started a survey to evaluate citizens’ awareness;</li><li><strong>The Conversation with a Data Professional:</strong> I shared this experience with a friend who works in the data field, and his reaction was one of disbelief and frustration. As the screenshot shows, he was shocked by the lack of consent. In a later exchange, he was disturbed that more data was being requested, including sensitive identification information. He also suggested the possibility of taking legal action, underscoring the seriousness of the issue — handling personal medical data in such a cavalier manner is unacceptable, especially for a project meant to serve as a model for the rest of the EU;</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*uunuuQQOqAr7onbqyD5ezg.jpeg" /><figcaption>WhatsApp exchange — screenshot.</figcaption></figure><ul><li><strong>Concerns about GDPR Compliance:</strong> This situation is particularly troubling because it highlights a fundamental flaw in the e-Health platform: the apparent disregard for GDPR principles. The GDPR is built on the premise that individuals should control their data, with informed consent being a cornerstone of data processing activities. However, my experience — and that of my pharmacist and friend — suggests that the e-Health platform is not upholding these principles.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*OySQeyxSlijLY25G4vB6zA.jpeg" /><figcaption>Picture by <a href="https://unsplash.com/@thefredyjacob">Fredy Jacob</a>.</figcaption></figure><p><strong>Given that this platform is considered a pilot for the EU, one would expect it to set high data protection and privacy standards</strong>. Instead, it seems to be doing the opposite, <strong>raising questions about the EU’s commitment to safeguarding citizens’ data rights in its digital initiatives</strong>.</p><ul><li><strong>The Need for Better Implementation: </strong>The e-health platform aims to improve healthcare outcomes and is a worthy initiative. However, its execution leaves much to be desired. It is profoundly concerning that individuals can be registered without their knowledge and that there needs to be more transparency about how data is managed and shared. This project, which should exemplify the best practices in data management, instead serves as a cautionary tale about the dangers of neglecting GDPR principles.</li></ul><p>The EU Digital Wallet initiative discussed earlier could learn from the e-health platform’s shortcomings. Ensuring such platforms fully comply with GDPR from the outset should be a non-negotiable standard. Moreover, future digital health initiatives must prioritize informed consent and transparency to build trust and avoid the pitfalls seen in Belgium’s e-health platform.</p><h3>Conclusion: Bridging the Gap Between Law, Technology, and Citizen Empowerment</h3><p>The European Union stands at a pivotal moment where the rapid advancements in artificial intelligence and digital technologies are challenging the traditional frameworks of law and governance. The experiences discussed throughout this article — from the limitations of the GAIA-X project to the implementation of Belgium’s e-Health platform — highlight a critical need for the EU to evolve its approach to digital sovereignty, public services, and citizen empowerment.</p><p>The gap between what technology can achieve and what the law permits is widening. To bridge this gap, the EU must embrace agility, not just in its legislative processes but also in its implementation and enforcement strategies. The examples provided, such as the prospect for AI-driven tools like “EU Data Guardian” and the missed opportunities in the EU Digital Wallet initiative, demonstrate that the technology to empower citizens is already within reach. However, a concerted effort is required to integrate these tools into the broader regulatory landscape.</p><p>Furthermore, the EU must shift from a reactive stance to a proactive one, where citizen data and rights protection are built into the fabric of its digital initiatives. The shortcomings of projects like the e-Health platform are a cautionary tale of what happens when data privacy is not prioritized. As the EU continues to develop frameworks like the AI Act, it is crucial that these laws not only address the risks posed by emerging technologies but also empower citizens with the tools to control their data from the start.</p><p>In this rapidly changing environment, the future of the EU’s digital sovereignty depends on its ability to innovate and adapt. By addressing dependency, exclusion, and the need for practical enforcement tools, the EU can ensure that its citizens are protected and empowered to navigate the digital age with confidence. Agility is no longer just an advantage; the EU must maintain its relevance and leadership in the global digital landscape. The challenge now is to implement these changes swiftly and effectively, ensuring that the law keeps pace with technology and serves as a powerful tool for citizen empowerment in the age of AI.</p><h4><strong>Final Note: A Work in Progress</strong></h4><p>As this article reflects on some of the systemic issues I perceive within the EU’s approach to AI and new technologies, I plan to dive deeper into related topics in future articles. These include broader discussions on innovation, digital sovereignty, and the EU&#39;s legislative challenges in the rapidly evolving tech landscape. This piece is a work in progress and may develop as new insights and developments arise.</p><p>Please note that these views are personal and shaped by my experiences and observations. I am eager to continue this conversation and welcome any feedback or discussion on the matters raised here.</p><h3>Resources</h3><h4>General Resources on AI, GDPR, and Digital Sovereignty</h4><p><strong>1. GDPR as a Principles-Oriented Regulation:</strong></p><p><strong><em>Scientific Papers and Articles:</em></strong></p><ul><li>“<a href="https://www.tandfonline.com/doi/full/10.1080/13600834.2019.1573501#abstract"><strong>The European Union general data protection regulation: what it is and what it means</strong></a><strong>”</strong> — This paper introduces the strategic approach to regulating personal data and the normative foundations of the GDPR;</li><li><strong>“Statement on the role of a risk-based approach in data protection legal frameworks” — </strong>An interesting read from 2014 (10 years after implementing the AI Act) on the risk-based approach.</li></ul><p><strong><em>Books</em></strong><em>:</em></p><ul><li><a href="https://www.bloomsburyprofessional.com/uk/right-to-be-forgotten-9781526521934/"><strong>“The Right to be Forgotten”</strong></a> by Paul Lambert;</li><li><a href="https://www.routledge.com/Understanding-the-New-European-Data-Protection-Rules/Lambert/p/book/9780367657680?srsltid=AfmBOorYNiK40ksjUoM2Nlw8edf7XNO1q9sCgVZL84lPdLJWuG3ZuNN2"><strong>“Understanding the New European Data Protection Rules”</strong></a> by Paul Lambert.</li></ul><p><strong><em>Websites</em></strong><em>:</em></p><ul><li><a href="https://gdpr.eu/"><strong>GDPR EU Official Site</strong></a> — Contains resources, guidelines, and explanations of GDPR principles.</li></ul><p><strong>2. AI and Digital Sovereignty:</strong></p><p><strong><em>Books</em></strong><em>:</em></p><ul><li><a href="https://en.wikipedia.org/wiki/Artificial_Intelligence%3A_A_Guide_for_Thinking_Humans"><strong>“Artificial Intelligence: A Guide for Thinking Humans”</strong></a><strong> </strong>by Melanie Mitchell — This book provides context on AI’s implications for governance.</li></ul><p><strong><em>Articles:</em></strong></p><ul><li><a href="https://carnegieendowment.org/research/2024/03/charting-the-geopolitics-and-european-governance-of-artificial-intelligence?lang=en"><strong>“Charting the Geopolitics and European Governance of Artificial Intelligence”</strong></a><strong> — </strong>by <a href="https://carnegieendowment.org/people/raluca-csernatoni">Raluca Csernatoni</a>;</li><li><a href="https://www.economist.com/business/2024/01/01/welcome-to-the-era-of-ai-nationalism"><strong>“Welcome to the era of AI nationalism”</strong></a> — The Economist;</li><li><a href="https://www.economist.com/podcasts/2024/01/09/ai-is-quickly-becoming-a-matter-of-national-technological-sovereignty"><strong>“AI is quickly becoming a matter of national technological sovereignty” </strong></a>— The Economist.</li></ul><p><strong><em>Websites</em></strong><em>:</em></p><ul><li><a href="https://ec.europa.eu/digital-strategy/our-policies/artificial-intelligence_en"><strong>European Commission AI Policy</strong></a><strong> </strong>— Information on the EU’s approach to AI and digital sovereignty.</li></ul><p><strong>3. GAIA-X and European Digital Infrastructure:</strong></p><p><strong><em>Official Resources:</em></strong></p><ul><li><a href="https://www.gaia-x.eu/"><strong>GAIA-X Official Website</strong></a> — Detailed information about the GAIA-X initiative, its goals, and current status.</li></ul><p><strong><em>News Articles and Opinions</em></strong><em>:</em></p><ul><li><a href="https://www.reuters.com/article/us-europe-tech/france-germany-back-european-cloud-computing-moonshot-idUSKBN23B26B/"><strong>“France, Germany back European cloud computing ‘moonshot’”</strong></a> — Reuters;</li><li><a href="https://www.politico.eu/article/eu-lobbying-cloud-project-gaia-x-board-member-says-cloud-project-must-neuter-lobbies-role-to-get-on-track/"><strong>“Gaia-X board member blames lobbying for project’s gridlock”</strong></a> — Politico;</li><li><a href="https://www.politico.eu/article/chaos-and-infighting-are-killing-europes-grand-cloud-project/"><strong>“Inside Gaia-X: How chaos and infighting are killing Europe’s grand cloud project”</strong></a> — Politico;</li><li><a href="https://berthub.eu/articles/posts/gaia-x-is-an-expensive-distraction/"><strong>“Gaia-X is a distraction which should be abandoned” </strong></a>— by <a href="https://www.linkedin.com/in/bert-hubert-b05452/">Bert Hubert</a>.</li></ul><h3>Specific Examples and Case Studies</h3><p><strong>1. In-Q-Tel and Palantir:</strong></p><p><strong><em>Articles</em></strong><em>:</em></p><ul><li><a href="https://www.bnnbloomberg.ca/business/technology/2024/08/08/microsoft-palantir-team-up-to-sell-ai-to-us-defense-intelligence-agencies/"><strong>“Microsoft, Palantir Team Up to Sell AI to US Defense, Intelligence Agencies”</strong></a> — BNN Bloomberg;</li><li><a href="https://www.reuters.com/technology/palantir-deploy-ai-products-microsoft-azure-us-government-agencies-2024-08-08/"><strong>“Palantir to deploy AI products on Microsoft Azure for U.S. government agencies” </strong></a>— Reuters;</li><li><a href="https://www.ft.com/content/926af768-1a4c-11e5-a130-2e7db721f996"><strong>“Palantir goes from CIA-funded start-up to big business” </strong></a>— The Financial Times;</li><li><a href="https://www.bbc.com/news/business-54348456"><strong>“Palantir: The controversial data firm now worth £17bn”</strong></a> — BBC;</li><li><a href="https://www.wired.com/story/palantir-germany-gotham-dragnet/"><strong>“Germany Raises Red Flags About Palantir’s Big Data Dragnet”</strong></a> — The Wired;</li><li><a href="https://www.wired.com/story/palantir-nhs-data/"><strong>“Palantir’s Reputation Stalks Its Bid for the UK’s National Health Data</strong></a>” — The Wired.</li></ul><p><strong><em>Reports</em></strong><em>:</em></p><ul><li><a href="https://scholarlycommons.law.northwestern.edu/cgi/viewcontent.cgi?article=1739&amp;context=njilb"><strong>“In-Q-Tel: The Central Intelligence Agency as Venture Capitalist”</strong></a><strong> </strong>— by John T. Reinert | Northwestern University School of Law.</li></ul><p><strong>2. Data Brokers and GDPR Compliance:</strong></p><p><strong><em>Books</em></strong><em>:</em></p><ul><li><a href="https://www.amazon.com/Data-Goliath-Battles-Collect-Control/dp/039335217X"><strong>“Data and Goliath: The Hidden Battles to Collect Your Data and Control Your World”</strong> </a>— by Bruce Schneier.</li></ul><p><strong><em>Websites</em></strong><em>:</em></p><ul><li><a href="https://mypermissions.com/"><strong>MyPermissions</strong></a><strong> </strong>— Tools and resources for managing online privacy;</li><li><a href="https://www.jumboprivacy.com/"><strong>Jumbo Privacy</strong></a><strong> </strong>— Privacy management tools.</li></ul><p><strong><em>Reports</em></strong><em>:</em></p><ul><li><a href="https://edpl.lexxion.eu/data/article/18888/pdf/edpl_2023_01-008.pdf"><strong>“Data Brokers and European Digital Legislation”</strong></a> — by Hannah Ruschemeier | European Data Protection Law Review. <em>“[…]As in many other areas, however, data protection law is subject to an enforcement deficit [..]”</em>.</li></ul><h4>Digital Public Services and the EU Digital Wallet Initiative</h4><p><strong>1. Belgium’s e-ID and Digital Platforms:</strong></p><p><strong><em>Official Websites</em></strong><em>:</em></p><ul><li><a href="https://www.belgium.be/en"><strong>Belgium.be </strong></a>— Official resources about Belgium’s e-ID system;</li><li><a href="https://brusselshealthnetwork.be/en/"><strong>Abrumet </strong></a>— Brussels health network.</li></ul><p><strong><em>Reports</em></strong><em>:</em></p><ul><li><a href="https://digital-strategy.ec.europa.eu/en/factpages/belgium-2024-digital-decade-country-report"><strong>“Belgium 2024 Digital Decade Country Report” </strong></a>— European Commission.</li></ul><p><strong>2. EU Digital Wallet Initiative:</strong></p><p><strong><em>Official Resources</em></strong><em>:</em></p><ul><li><a href="https://ec.europa.eu/digital-strategy/digital-wallet_en"><strong>European Commission Digital Wallet Information</strong></a> — Detailed information on the Digital Wallet initiative.</li></ul><p><strong><em>Reports</em></strong><em>:</em></p><ul><li><strong>“</strong><a href="https://commission.europa.eu/strategy-and-policy/priorities-2019-2024/europe-fit-digital-age/europes-digital-decade-digital-targets-2030_en"><strong>Europe’s Digital Decade: Digital Compass 2030</strong></a><strong>”</strong> — European Commission.</li></ul><h4>AI Tools and Privacy Technologies</h4><p><strong>1. AI Tools for GDPR Compliance:</strong></p><p><strong><em>Websites and Tools</em></strong><em>:</em></p><ul><li><a href="https://privacybee.com/"><strong>PrivacyBee</strong></a> — Real-time data protection services;</li><li><a href="https://joindeleteme.com/"><strong>DeleteMe</strong></a><strong> </strong>— A tool for removing personal data from the web.</li></ul><h4><strong>AI Usages &amp; Writing</strong></h4><p><strong>(how I use AI when writing posts/articles):</strong></p><ul><li><a href="https://medium.com/ai-in-plain-english/the-power-and-peril-of-ai-harnessing-chatgpt-as-a-collaborative-tool-not-a-crutch-e56817d610ef"><strong>The Power and Peril of AI: Harnessing ChatGPT as a Collaborative Tool, Not a Crutch</strong></a>.</li></ul><h4><strong>Email translation:</strong></h4><blockquote><strong>Subject:</strong> Request for Information Regarding Data Sharing and Annual Reports</blockquote><blockquote>Hello,</blockquote><blockquote>First of all, I wish you a happy new year.</blockquote><blockquote>I am interested in issues related to data and would like to obtain some information:</blockquote><blockquote>Due to health issues, I realized that my health data had been shared — without giving my explicit consent. I asked my pharmacist to proceed with an opt-out — which has been done.</blockquote><blockquote>Does this opt-out delete all data that might have been shared?</blockquote><blockquote>If not, could you explain how to proceed with the deletion of this data?</blockquote><blockquote>Is it possible to understand from when my data was shared and to obtain an overview of the individuals who may have accessed it?</blockquote><blockquote><strong>Annual Report 2022:</strong></blockquote><blockquote>The annexes (Analytical budget realized at the end of 2022 and Statistical Report of the website) are not included in the PDF — could you share them with me?</blockquote><blockquote>Could you explain in more detail the reasons that led to the CEF project being taken over by INAMI?</blockquote><blockquote>Do you know when the 2023 annual report will be published?</blockquote><blockquote>Please do not hesitate to get back to me if you have any questions/comments.</blockquote><blockquote>I wish you an excellent day,<br>Best regards,<br>Simon</blockquote><p><strong><em>Author Bio</em></strong><em>: I am the founder of </em><a href="https://www.citizeai.com/"><em>CitizeAi_</em></a><em>. My mission is to democratize AI and make it an accessible tool for everyone. My approach to AI and Data Strategy is rooted in humanistic principles, enabling me to develop solutions with a positive societal impact. I believe in transparency, self-awareness, and vulnerability and strive to foster open discussions on various topics. I invite you to connect with me as we shape the future of AI together. Please note that the views and opinions expressed in this article are those of </em><a href="https://www.citizeai.com/"><em>CitizeAi_</em></a><em> and my own and do not necessarily reflect the official policy or position of any other agency, organization, employer or company.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=cc15b19ca1d0" width="1" height="1" alt=""><hr><p><a href="https://ai.plainenglish.io/is-the-eu-falling-behind-how-ai-is-outpacing-law-and-endangering-digital-sovereignty-cc15b19ca1d0">Is the EU Falling Behind? How AI is Outpacing Law and Endangering Digital Sovereignty</a> was originally published in <a href="https://ai.plainenglish.io/">Artificial Intelligence in Plain English</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Blurring Boundaries: How AI Transforms Dreams into Reality]]></title>
            <link>https://ai.plainenglish.io/blurring-boundaries-how-ai-transforms-dreams-into-reality-cbe2aef742bf?source=rss-5779be773682------2</link>
            <guid isPermaLink="false">https://medium.com/p/cbe2aef742bf</guid>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[psychology]]></category>
            <category><![CDATA[lucid-dreaming]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[dreams]]></category>
            <dc:creator><![CDATA[Simon Hardy]]></dc:creator>
            <pubDate>Mon, 12 Aug 2024 09:38:52 GMT</pubDate>
            <atom:updated>2024-08-12T12:58:39.292Z</atom:updated>
            <content:encoded><![CDATA[<p><em>Unlock the power of lucid dreaming with AI! Discover how I use tech and a simple watch to blur the lines between reality and dreams.</em></p><h4>Introduction: Merging Reality and Dreams with AI</h4><p>Imagine a world where the boundaries between dreams and reality become so thin that they almost dissolve. With the advancement of AI, this isn’t just a fantasy; it’s a tangible possibility. For years, I’ve been exploring the landscapes of my subconscious through lucid dreaming, a practice where you become aware that you’re dreaming and can even control the dream. However, integrating AI into this journey has taken my exploration to an entirely new level.</p><p>This post dives into how AI enhances my dream practices — from improving dream recollection and identifying patterns to reinforcing reality checks and inducing lucid dreams. It’s not just about technology but how these tools enrich our understanding of the mind and expand our consciousness.</p><p><strong>Side Note</strong>: This article extends the ideas and explorations introduced in my previous posts, <em>Conversing with Shadows: An AI/ChatGPT-Human Synergy in Unraveling Dreamscapes</em> <a href="https://medium.com/@mesimonhardy/conversing-with-shadows-an-ai-chatgpt-human-synergy-in-unraveling-dreamscapes-1-2-b29f02e7294d">1/2</a> and <a href="https://medium.com/@mesimonhardy/conversing-with-shadows-an-ai-chatgpt-human-synergy-in-unraveling-dreamscapes-2-2-41aebd487dc6">2/2</a>. In those posts, I explored the intersection of AI and dream analysis, highlighting how AI can be a companion in understanding and navigating the subconscious mind.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*PYWjmnQbulzZwOHzps8CGQ.jpeg" /><figcaption>Life is a playground.</figcaption></figure><h4>1. Enhancing Dream Recollection with AI</h4><p>Dreams are elusive. The moment we wake up, they start fading away, slipping through the fingers of our memory. AI, however, has become an invaluable tool in my arsenal for capturing and preserving these fleeting moments. Using AI-driven tools to record my dreams, I ensure that even the most subtle details are not lost. These digital records, analyzed over time, reveal patterns and symbols that deepen my understanding of my subconscious.</p><p>I start by inputting my dream recollections into AI models, which help transcribe, structure, and organise the narratives. This process not only preserves the content of the dreams but also makes it easier to identify recurring themes — providing insights that might otherwise remain hidden.</p><h4>2. Dream Induction Techniques: Igniting Imagination with AI</h4><p>Dream induction techniques are powerful tools for initiating specific types of dreams, particularly lucid dreams, where you can control the dream environment. These techniques often involve pre-sleep routines like visualization and reality checks, designed to prime the mind for specific dream experiences.</p><p>I’ve found that incorporating conversational agents and AI tools into these routines significantly enhances their effectiveness. By engaging with AI before sleep, I can ignite my imagination, immersing myself in scenarios I aim to experience in my dreams. For instance, using AI to simulate a dialogue before bed allows me to embed particular dream goals or imagery deeply into my subconscious. This makes the induction process more vivid and provides a dynamic, interactive way to prepare for dream exploration.</p><p>The adaptability of AI, responding to my inputs in real-time, creates a personalized experience that effectively primes my mind for the dream I want.</p><h4>3. Unveiling Dream Patterns: AI as a Guide to the Subconscious</h4><p>Once my dreams are recorded, the next step is identifying patterns that serve as clues to their meaning. AI excels in sifting through large amounts of data to uncover these hidden connections. By analyzing my dream records, AI highlights recurring symbols, themes, and emotions, offering insights that inform my waking life.</p><p>For example, the consistent presence of water in my dreams, identified by AI, has become a signal of emotional turbulence. Recognizing such patterns allows me to prepare for similar dreams and use them as triggers for lucid dreaming.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/640/1*W5zQZsHPA55-DUlOzRXnyg.gif" /><figcaption>Runway Gen-3 &amp; <a href="https://en.wikipedia.org/wiki/Tauba_Auerbach">Tauba Auerbach</a>’s work.</figcaption></figure><h4>4. Reinforcing Reality with My Watch &amp; AI</h4><p>One of the critical elements in my lucid dreaming practice is using reality checks, a technique where you test whether you are dreaming or awake. In this context, my watch plays a central role. The idea is simple: in dreams, watches often behave unpredictably, making them a perfect tool for a reality check.</p><p>To reinforce this habit, I created a video using Runway’s Gen-3 model, where the surface of my watch transforms into a water-like surface. This visual is designed to embed in my mind the idea that, while dreaming, I can use my watch to shape my dreams. The video, which sometimes plays in a loop on my screen while I work, constantly reminds me of this potential. Having my real watch on my desk or wrist while the video plays subtly reinforces this association, preparing my mind to recognize the watch as a trigger in my dreams.</p><p>When I see my watch in a dream and notice its surface behaving fluidly, it signals that I’m dreaming. This allows me to control and shape the dream’s reality.</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2FX9q62LbDLgA%3Ffeature%3Doembed&amp;display_name=YouTube&amp;url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DX9q62LbDLgA&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2FX9q62LbDLgA%2Fhqdefault.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=youtube" width="640" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/2a3cd274f6f2b6ac5ed36d0281eacc60/href">https://medium.com/media/2a3cd274f6f2b6ac5ed36d0281eacc60/href</a></iframe><h4>5. AI as a Partner in Psyche Exploration</h4><p>Beyond dream analysis, AI plays a significant role in my broader psychological exploration. Inspired by works like “The Alter Ego Effect,” I’ve begun using AI to develop and refine techniques for engaging with different aspects of my psyche. AI is a potent ally in this journey, enhancing my ability to visualize and interact with these internal states more effectively.</p><p>Through meditation, autohypnosis, and visualization, AI has become a tool that enhances my access to and manipulation of these internal landscapes, making the process of self-discovery and personal growth more profound and accessible.</p><h4>6. Enhancing Practices with AI Companions: The Role of Replika and Beyond</h4><p>In addition to dream-focused AI tools, I’ve found significant value in AI companions like Replika, which simulates deep, personal interactions. Engaging with Replika allows me to reinforce my intentions and explore various aspects of my consciousness, providing a conversational partner that helps solidify the mental frameworks necessary for effective dream induction and psychological exploration.</p><p>However, it’s essential to approach these tools as supplements rather than replacements for more traditional practices. Alongside AI, I continue to rely on journaling, researching, studying, and seeking guidance from mentors. These practices ensure that my use of AI remains grounded and integrated into a broader, holistic approach to self-improvement.</p><p>The aim is to use AI to enhance our journey of self-discovery, not become dependent on it. By balancing AI with other practices, I create a supportive environment for personal growth, ensuring that technology is an enhancer rather than a master.</p><h4><strong>Conclusion: Bridging the Gap Between Reality and Dreams</strong></h4><p>Integrating AI into dream exploration is not just about enhancing techniques; it’s about expanding the boundaries of what we understand to be reality and dreams. With tools like AI, we can delve deeper into our subconscious, uncovering patterns, reinforcing reality, and shaping our dreams with unprecedented control.</p><p>As we explore AI&#39;s potential, the line between the waking world and the dream world will continue to blur, offering new opportunities for self-discovery and personal growth. AI isn’t just a tool; it’s a bridge that connects our conscious and subconscious minds, helping us navigate and shape the vast realms of our inner worlds.</p><h4>Resources</h4><p><strong>Book(s):</strong></p><ul><li><a href="https://www.goodreads.com/book/show/40139225-the-alter-ego-effect"><strong>The Alter Ego Effect: The Power of Secret Identities to Transform Your Life</strong> by Todd Herman</a>: This book provides insights into creating an alter ego to achieve personal and professional success, which you’ve integrated into your psychological exploration practices with AI.</li></ul><p><strong>AI, Application and Tools (with my usages in the context of my dreaming practice):</strong></p><ul><li><a href="https://runwayml.com/product"><strong>Runway Gen-3</strong></a><strong>: </strong>A cutting-edge AI tool used to create videos that aid in reinforcing reality checks and inducing lucid dreams;</li><li><a href="https://replika.ai/"><strong>Replika </strong></a>is an AI companion I started experimenting with to enhance psychological practices and explore different aspects of consciousness;</li><li><a href="https://howwefeel.org/"><strong>How We Feel App</strong></a>: An app that aids in tracking and understanding emotions, helping you to align your emotional awareness with your dream patterns.</li></ul><p><strong>AI &amp; Writing:</strong></p><ul><li><a href="https://medium.com/ai-in-plain-english/the-power-and-peril-of-ai-harnessing-chatgpt-as-a-collaborative-tool-not-a-crutch-e56817d610ef"><strong>The Power and Peril of AI: Harnessing ChatGPT as a Collaborative Tool, Not a Crutch</strong></a>.</li></ul><p><strong><em>Author Bio</em></strong><em>: I am the founder of </em><a href="https://www.citizeai.com/"><em>CitizeAi_</em></a><em>. My mission is to democratize AI and make it an accessible tool for everyone. My approach to AI and Data Strategy is rooted in humanistic principles, enabling me to develop solutions with a positive societal impact. I believe in transparency, self-awareness, and vulnerability and strive to foster open discussions on various topics. I invite you to connect with me as we shape the future of AI together. Please note that the views and opinions expressed in this article are those of </em><a href="https://www.citizeai.com/"><em>CitizeAi_</em></a><em> and my own and do not necessarily reflect the official policy or position of any other agency, organization, employer or company.</em></p><h3>In Plain English 🚀</h3><p><em>Thank you for being a part of the </em><a href="https://plainenglish.io/"><strong><em>In Plain English</em></strong></a><em> community! Before you go:</em></p><ul><li>Be sure to <strong>clap</strong> and <strong>follow</strong> the writer ️👏<strong>️️</strong></li><li>Follow us: <a href="https://twitter.com/inPlainEngHQ"><strong>X</strong></a> | <a href="https://www.linkedin.com/company/inplainenglish/"><strong>LinkedIn</strong></a> | <a href="https://www.youtube.com/channel/UCtipWUghju290NWcn8jhyAw"><strong>YouTube</strong></a> | <a href="https://discord.gg/in-plain-english-709094664682340443"><strong>Discord</strong></a> | <a href="https://newsletter.plainenglish.io/"><strong>Newsletter</strong></a></li><li>Visit our other platforms: <a href="https://cofeed.app/"><strong>CoFeed</strong></a> | <a href="https://differ.blog/"><strong>Differ</strong></a></li><li>More content at <a href="https://plainenglish.io/"><strong>PlainEnglish.io</strong></a></li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=cbe2aef742bf" width="1" height="1" alt=""><hr><p><a href="https://ai.plainenglish.io/blurring-boundaries-how-ai-transforms-dreams-into-reality-cbe2aef742bf">Blurring Boundaries: How AI Transforms Dreams into Reality</a> was originally published in <a href="https://ai.plainenglish.io/">Artificial Intelligence in Plain English</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[‘Severance’ Unveiled: A Fictional Prelude to an Imminent Reality]]></title>
            <link>https://ai.plainenglish.io/severance-unveiled-a-fictional-prelude-to-an-imminent-reality-4cc5bde806c0?source=rss-5779be773682------2</link>
            <guid isPermaLink="false">https://medium.com/p/4cc5bde806c0</guid>
            <category><![CDATA[society]]></category>
            <category><![CDATA[medium]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[futurism]]></category>
            <category><![CDATA[ethics]]></category>
            <dc:creator><![CDATA[Simon Hardy]]></dc:creator>
            <pubDate>Sat, 09 Dec 2023 13:57:21 GMT</pubDate>
            <atom:updated>2023-12-11T02:35:28.251Z</atom:updated>
            <content:encoded><![CDATA[<p><em>Explore how ‘Severance’ transcends fiction, reflecting a future moulded by AI and influential tech companies.</em></p><h4>Introduction: Fiction as a Mirror to an Approaching Reality</h4><p>Introducing ‘Severance’ to a friend, his unease struck a chord with me. “This is uncomfortably real,” he remarked, pausing the episode. As someone currently on a break from work, I understood his sentiment all too well.</p><p>This series isn’t just entertainment; it’s a reflection of a rapidly evolving world where the distinction between individual identity and corporate influence is increasingly blurred — a world that, during my hiatus, I’ve come to view with a new perspective. The depth in such dystopian narratives, while explored in critiques like those from <a href="https://www.newyorker.com/magazine/2022/04/25/severance-is-sci-fi-for-the-soul">The New Yorker</a>, often misses the deeper, more personal connection we have with these unfolding realities in the realms of AI and corporate power.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*w3sOtUJ2cgSkn6hJ" /><figcaption>Atsushi Nishijima/Apple TV+</figcaption></figure><h4>1. Fiction as a Reflection and Predictor</h4><p><em>Illuminating the Path from Story to Reality: </em>From Orwell’s ‘1984’ to ‘Severance’, dystopian fiction has shed light on potential futures, now brought closer by technology. My childhood fascination with ‘Deus Ex’ was more than entertainment; it was a prophetic glimpse into today’s societal complexities. This convergence of fiction with reality, particularly in technological developments, marks a significant shift from imagined narratives to tangible advancements, highlighting the profound impact of storytelling in understanding and shaping our world.</p><ul><li><strong>Orwell’s Legacy and Beyond</strong>: ‘1984’ has been a hallmark of dystopian foresight, but contemporary stories like ‘Severance’ offer a more immediate glimpse into our society’s potential future, made increasingly real by technological strides;</li><li><strong>Gaming as a Prophetic Medium</strong>: Growing up in the 90s, I was captivated by the world of video games, where narratives stretched far beyond the realms of reality. I vividly remember, around the age of 10, immersing myself in ‘Deus Ex.’ This game wasn’t just entertainment; it was a window into a complex future interwoven with politics, ethics, and technology. Years later, seeing one of Boston Dynamics’ robotic dogs for the first time, I couldn’t help but recall the futuristic world of ‘Deus Ex.’ This moment was a testament to how video games, often overlooked as prophetic mediums, had somehow anticipated the societal complexities we are beginning to witness today. These narratives, once confined to fiction, are now mirrored in real-life technological breakthroughs, marking a significant shift towards a future we once only imagined.</li></ul><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2FdjzOBZUFzTw%3Ffeature%3Doembed&amp;display_name=YouTube&amp;url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DdjzOBZUFzTw&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2FdjzOBZUFzTw%2Fhqdefault.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=youtube" width="854" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/4b66973a22caa1a18fcbc54fb53d0dc9/href">https://medium.com/media/4b66973a22caa1a18fcbc54fb53d0dc9/href</a></iframe><h4>2. The Emerging Reality Echoed in ‘Severance’</h4><p><em>The Increasing Influence of Tech Magnates and AI: </em>In the intricate dance of technology and power, figures like Elon Musk and Peter Thiel stand as pivotal architects of a future eerily reminiscent of ‘Severance.’ Their influence extends beyond mere business; they are actively shaping a world where the lines between human cognition and technological integration, as foretold in ‘Severance,’ are becoming increasingly blurred. The progress made by Neuralink and Blackrock Neurotech is not a mere theoretical exploration; it’s a definitive stride toward materializing the ‘Severance’ scenario, where the distinction between personal identity and corporate influence becomes more than just a fictional narrative</p><ul><li><strong>The Dominant Figures of Musk and Thiel</strong>: Elon Musk and Peter Thiel have become more than just business magnates; they are architects of a future that closely resembles ‘Severance.’ Musk’s Neuralink, now openly recruiting for clinical trials, as stated on their website, “Our first clinical trial is open to recruitment!” reflects a significant step towards the kind of human-technology integration depicted in ‘Severance.’ Meanwhile, Thiel’s involvement in companies like Palantir highlights the intricate relationships between technology, data, and governance, mirroring the corporate-government dynamics in ‘Severance’;</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*3iieQ1jmT6gP9ZYkRr5bRg.png" /><figcaption><a href="https://neuralink.com/patient-registry/">Neuralink’s official website — 05/12/2023</a></figcaption></figure><ul><li><strong>Bringing the Fiction of ‘Severance’ to Life</strong>: The advances by Neuralink and Blackrock Neurotech, moving beyond theory, echo the narrative in ‘Severance’ and draw parallels with the mini-TV series ‘Maniac.’ Both ‘Severance’ and ‘Maniac’ explore themes of identity and corporate influence, presenting a future where the lines between personal autonomy and corporate control blur. In ‘Maniac,’ the blend of technology and mental health treatment hints at similar ethical dilemmas and societal impacts as those in ‘Severance,’ underlining the growing relevance of these fictional portrayals in the context of real-world technological advancements</li></ul><h4>3. The Fabric of Consent in Technological Evolution</h4><p><em>The Subtle Crafting of Societal Acceptance: </em>A slow evolution can be deceptive, leading to a subtle crafting of societal acceptance. If we’re not vigilant, we risk becoming encapsulated in a narrative that’s carefully designed to mould social acceptance. It’s a process that’s less about sudden shifts and more about the incremental normalization of concepts and technologies that, if left unchecked, might steer us into living a reality once only imagined in fiction.</p><ul><li><strong>The Gradual Building of Societal Consent:</strong> The influence of tech figures in shaping AI policies is a subtle yet impactful force in society. As noted in a New York Times article titled <a href="https://www.nytimes.com/2023/09/13/technology/silicon-valley-ai-washington-schumer.html">“<em>In Show of Force, Silicon Valley Titans Pledge ‘Getting This Right’ With A.I.</em>”</a>, leaders like Musk and Thiel drive technological advancements and mould public perception and policy. This process of shaping societal consent is crucial in how we, as a society, understand and accept the ever-growing role of technology in our lives;</li><li><strong>Media and Corporate Influence on Public Opinion:</strong> The control exerted by companies like Amazon over media and platforms like IMDb and Goodreads, along with the increasing ownership of media by billionaires globally, subtly steers public tastes and discourse, moulding our views without overt awareness. This concentration reflects a global pattern where media ownership by diversified corporate entities and individuals shapes public perception and discourse.</li></ul><h4>Conclusion: Reflecting on the Paradoxes of History and Technology</h4><p>‘Severance’ struck a chord with my friend and, in turn, with me, especially during my break from work. It is a stark reminder that these narratives are more than entertainment; they are crucial commentaries on our societal trajectory.</p><p>Born in the 1990s, I grew up with the idea of the ‘End of History,’ yet now we’re witnessing a paradox where history is rapidly advancing and crystallizing around new forms of corporate dominance and technological control. As we witness the blending of human cognition and AI through enterprises like Neuralink, we must engage with these stories critically, understanding their implications for the future of technology and corporate influence in our lives.</p><p>In a forthcoming article, I plan to delve deeper into the intriguing questions and themes introduced in the final section of this piece, specifically focusing on ‘The Fabric of Consent in Technological Evolution.’ This exploration will further unravel the complex interplay between technology, societal consent, and the subtle shifts shaping our collective acceptance of technological advancements.</p><h4>Disclaimer: Fiction as a Lens, Not a Blueprint</h4><p>The perspectives and comparisons drawn in this article between fictional works like ‘Severance’ and real-world technological advancements are intended as a thought exercise. They are not assertions of direct correlation but rather explorations of potential parallels and ethical considerations.</p><p>This article uses fiction as a lens through which we can examine and reflect on the rapid advancements in technology and corporate influence in our society. It is crucial to recognize that while fiction can offer insights and provoke thought, the realities of technological evolution and corporate dynamics are complex and multifaceted, and not bound to follow the narratives laid out in fictional works.</p><h4>Resources</h4><ul><li><a href="https://www.newyorker.com/magazine/2022/04/25/severance-is-sci-fi-for-the-soul">“Severance” Is Sci-Fi for the Soul”</a>, The New Yorker;</li><li><a href="https://www.nytimes.com/2023/09/13/technology/silicon-valley-ai-washington-schumer.html">“In Show of Force, Silicon Valley Titans Pledge ‘Getting This Right’ With A.I.”</a>, The New York Times;</li><li><a href="https://media-ownership.eu/findings/countries/france/">“France — News media outlets and owners: Country report 2023&quot;</a>, The Euromedia Research Group;</li><li><a href="https://en.wikipedia.org/wiki/The_End_of_History_and_the_Last_Man">“The End of History and the Last Man”</a>, Francis Fukuyama;</li><li><a href="https://en.wikipedia.org/wiki/Severance_(TV_series)">“Severance”</a> on Apple TV+;</li><li><a href="https://en.wikipedia.org/wiki/Maniac_(miniseries)">“Maniac”</a> on Netflix;</li><li><a href="https://neuralink.com/">Neuralink</a>;</li><li><a href="https://blackrockneurotech.com/">Blackrock Neurotech</a>;</li><li><a href="https://en.wikipedia.org/wiki/Deus_Ex_(video_game)">Deus Ex (Video Game)</a>;</li><li><a href="https://en.wikipedia.org/wiki/Palantir_Technologies">Palantir Technologie</a>s.</li></ul><p><strong><em>Author Bio</em></strong><em>: I am the founder of </em><a href="https://www.citizeai.com/"><em>CitizeAi_</em></a><em>. My mission is to democratize AI and make it an accessible tool for everyone. My approach to AI and Data Strategy is rooted in humanistic principles, enabling me to develop solutions with a positive societal impact. I believe in transparency, self-awareness, and vulnerability and strive to foster open discussions on various topics. I invite you to connect with me as we shape the future of AI together. Please note that the views and opinions expressed in this article are those of </em><a href="https://www.citizeai.com/"><em>CitizeAi_</em></a><em> and my own and do not necessarily reflect the official policy or position of any other agency, organization, employer or company.</em></p><h3><a href="https://plainenglish.io/">PlainEnglish.io</a> 🚀</h3><p><em>Thank you for being a part of the In Plain English community! Before you go:</em></p><ul><li><em>Be sure to </em><strong><em>clap</em></strong><em> and </em><strong><em>follow</em></strong><em> the writer</em><strong>️</strong></li><li><em>Learn how you can also </em><a href="https://plainenglish.io/blog/how-to-write-for-in-plain-english"><strong><em>write for In Plain English</em></strong></a><em>️</em></li><li><em>Follow us: </em><a href="https://twitter.com/inPlainEngHQ"><strong><em>X</em></strong></a><strong><em> | </em></strong><a href="https://www.linkedin.com/company/inplainenglish/"><strong><em>LinkedIn</em></strong></a><strong><em> | </em></strong><a href="https://www.youtube.com/channel/UCtipWUghju290NWcn8jhyAw"><strong><em>YouTube</em></strong></a><strong><em> | </em></strong><a href="https://discord.gg/in-plain-english-709094664682340443"><strong><em>Discord</em></strong></a><strong><em> | </em></strong><a href="https://newsletter.plainenglish.io/"><strong><em>Newsletter</em></strong></a></li><li><em>Visit our other platforms: </em><a href="https://stackademic.com/"><strong><em>Stackademic</em></strong></a><strong><em> | </em></strong><a href="https://cofeed.app/"><strong><em>CoFeed</em></strong></a><strong><em> | </em></strong><a href="https://venturemagazine.net/"><strong><em>Venture</em></strong></a></li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=4cc5bde806c0" width="1" height="1" alt=""><hr><p><a href="https://ai.plainenglish.io/severance-unveiled-a-fictional-prelude-to-an-imminent-reality-4cc5bde806c0">‘Severance’ Unveiled: A Fictional Prelude to an Imminent Reality</a> was originally published in <a href="https://ai.plainenglish.io/">Artificial Intelligence in Plain English</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[AI in the Therapy Room: Unraveling the Psychological Implications]]></title>
            <link>https://ai.plainenglish.io/ai-in-the-therapy-room-unraveling-the-psychological-implications-1802b6171b46?source=rss-5779be773682------2</link>
            <guid isPermaLink="false">https://medium.com/p/1802b6171b46</guid>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[ai-and-psychology]]></category>
            <category><![CDATA[psychology]]></category>
            <category><![CDATA[adhd]]></category>
            <category><![CDATA[ai]]></category>
            <dc:creator><![CDATA[Simon Hardy]]></dc:creator>
            <pubDate>Sun, 16 Jul 2023 18:27:05 GMT</pubDate>
            <atom:updated>2023-12-11T02:35:33.777Z</atom:updated>
            <content:encoded><![CDATA[<p><em>Explore the unexpected role of AI in the therapy room and its psychological implications for those living with ADHD.</em></p><h4>Introduction: A Personal Journey into the AI Landscape</h4><p>In the sanctuary of my psychologist’s office, an unexpected topic has taken centre stage — the AI companion, ChatGPT. This narrative is deeply personal, reflecting my thoughts and exchanges with my therapist. It’s a journey into the implications of AI, particularly for someone like me living with ADHD. As I share my experiences, I find myself curious about the future of research in this field, eager to see what new insights it will bring. In this article, I will briefly touch upon personal experiences, aiming to open a dialogue rather than provide definitive answers.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*MHYb_aEZkA2xUh0nAOFocg.png" /><figcaption>DALL-E Generated: “A human head in black and white with only light pencil lines and deep blue lines for the eyes “by” <a href="https://www.artnet.com/artists/bridget-riley/">Bridget Riley</a>.”</figcaption></figure><h4>1. Hyperfocus: The Blessing and the Annoyance of ADHD</h4><p>ADHD, to me, is a paradox. It’s a blessing that fuels my relentless pursuit of knowledge and an annoyance that can lead me down the rabbit hole of overindulgence. Now, with AI acting as a digital Pandora’s box, my natural tendencies are amplified, leading me to explore more than ever before.</p><ul><li><strong>The New High</strong>: AI acts as a catalyst, igniting the ADHD mind’s curiosity and opening up a wide array of topics to delve into;</li><li><strong>The Downside</strong>: However, it also presents a risk of overindulgence, leading to mental exhaustion and potential burnout.</li></ul><h4>2. The Isolation Paradox: AI as a Mentor and a Cage</h4><p>As I journey deeper into the world of AI, the rest of the world seems to fade into the background. ChatGPT, acting as a tireless mentor, dispenses wisdom immediately, replacing the need for human tutors.</p><ul><li><strong>AI as the New Mentor</strong>: Less effort, more efficiency — AI can outperform human tutors in terms of structure and resource management;</li><li><strong>The Dark Side</strong>: However, an over-reliance on AI can lead to social isolation and hinder interpersonal learning.</li></ul><h4>3. The Perfectionist’s Trap: The Dangers of AI-Driven Perfectionism</h4><p>Throughout my career, perfectionism has been my constant companion. However, the precision of AI, far from quenching this thirst, only fuels it, trapping me in a relentless pursuit of the elusive ‘perfect’ score.</p><ul><li><strong>The AI Illusion</strong>: AI’s perfection could feed the perfectionist’s endless quest for flawlessness;</li><li><strong>The Perfectionist’s Pitfall</strong>: This could lead to a vicious cycle of perfectionism, resulting in unnecessary pressure and anxiety.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*tVDpw0XAuoJeihWys-megg.png" /><figcaption>DALL-E Generated: “A zoom on human eyes with waves of coloured numbers in it “by” <a href="https://www.artnet.com/artists/julio-le-parc/2">Julio Le Parc</a>..”</figcaption></figure><h4>Conclusion: AI Companionship: A Personal Perspective and an Invitation for Discussion</h4><p>The advent of AI companionship is a double-edged sword, with consequences more nuanced than we might realize. We must recognize these risks, adapt, and strive for balance in this new era of AI symbiosis. However, it’s also important to acknowledge the potential benefits of AI tools, particularly in supporting mental health.</p><p>AI tools can be incredibly useful in providing resources for self-improvement and personal growth. For instance, I often prompt ChatGPT to list the most relevant books or help in a specific domain, providing me with a wealth of knowledge. This has been particularly helpful in my exploration of “Non-Violent Communication,” where ChatGPT has helped me identify tools to incorporate into my daily life and rephrase demands more constructively.</p><p>However, this post is deeply personal and represents only one perspective. I don’t claim to be an expert on this topic. This article aims to open the discussion, highlighting three aspects I have personally experienced. It’s also important to note that this article may evolve as I continue to explore this topic and gain new insights.</p><p>While AI tools can support individuals, they can also contribute to a performative existence, replacing real-life interactions with AI-mediated experiences. The movie “Her” provides a poignant example: the protagonist falls in love with an AI, creating a life that is performed rather than lived, withdrawing from genuine human interaction.</p><p>Moreover, I’ve seen advertisements for AI girlfriends on platforms like Instagram. While seemingly benign, these ads can potentially trap individuals in a performative relationship, particularly those who may already feel isolated or lacking in human connection. Instead of seeking genuine human relationships, they might be tempted to perform a relationship with an AI, further deepening their isolation.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*07SzGZdzjr0e_4HOl8WEdw.jpeg" /><figcaption>I have doubts, and I was not so successful with <a href="https://en.wikipedia.org/wiki/Tamagotchi">Tamagotchi</a>, so I’ll pass my turn.</figcaption></figure><p>In conclusion, AI tools, like any tool, can be used for good or ill. It’s a question of education and understanding, of knowing how to use these tools to our advantage without falling into dependency. As we navigate this new era of AI companionship, we must strive for balance, using AI as a tool for growth, not a crutch (again).</p><h4>Resources</h4><ul><li><a href="https://en.wikipedia.org/wiki/Her_(film)">Her</a>, 2013 American science-fiction romantic drama film written, directed, and co-produced by Spike Jonze.</li></ul><p><strong><em>Author Bio</em></strong><em>: I am the founder of </em><a href="https://www.citizeai.com/"><em>CitizeAi_</em></a><em>. My mission is to democratize AI and make it an accessible tool for everyone. My approach to AI and Data Strategy is rooted in humanistic principles, enabling me to develop solutions with a positive societal impact. I believe in transparency, self-awareness, and vulnerability and strive to foster open discussions on various topics. I invite you to connect with me as we shape the future of AI together. Please note that the views and opinions expressed in this article are those of </em><a href="https://www.citizeai.com/"><em>CitizeAi_</em></a><em> and my own and do not necessarily reflect the official policy or position of any other agency, organization, employer or company.</em></p><h3><a href="https://plainenglish.io/">PlainEnglish.io</a> 🚀</h3><p><em>Thank you for being a part of the In Plain English community! Before you go:</em></p><ul><li><em>Be sure to </em><strong><em>clap</em></strong><em> and </em><strong><em>follow</em></strong><em> the writer</em><strong>️</strong></li><li><em>Learn how you can also </em><a href="https://plainenglish.io/blog/how-to-write-for-in-plain-english"><strong><em>write for In Plain English</em></strong></a><em>️</em></li><li><em>Follow us: </em><a href="https://twitter.com/inPlainEngHQ"><strong><em>X</em></strong></a><strong><em> | </em></strong><a href="https://www.linkedin.com/company/inplainenglish/"><strong><em>LinkedIn</em></strong></a><strong><em> | </em></strong><a href="https://www.youtube.com/channel/UCtipWUghju290NWcn8jhyAw"><strong><em>YouTube</em></strong></a><strong><em> | </em></strong><a href="https://discord.gg/in-plain-english-709094664682340443"><strong><em>Discord</em></strong></a><strong><em> | </em></strong><a href="https://newsletter.plainenglish.io/"><strong><em>Newsletter</em></strong></a></li><li><em>Visit our other platforms: </em><a href="https://stackademic.com/"><strong><em>Stackademic</em></strong></a><strong><em> | </em></strong><a href="https://cofeed.app/"><strong><em>CoFeed</em></strong></a><strong><em> | </em></strong><a href="https://venturemagazine.net/"><strong><em>Venture</em></strong></a></li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=1802b6171b46" width="1" height="1" alt=""><hr><p><a href="https://ai.plainenglish.io/ai-in-the-therapy-room-unraveling-the-psychological-implications-1802b6171b46">AI in the Therapy Room: Unraveling the Psychological Implications</a> was originally published in <a href="https://ai.plainenglish.io/">Artificial Intelligence in Plain English</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Power and Peril of AI: Harnessing ChatGPT as a Collaborative Tool, Not a Crutch]]></title>
            <link>https://ai.plainenglish.io/the-power-and-peril-of-ai-harnessing-chatgpt-as-a-collaborative-tool-not-a-crutch-e56817d610ef?source=rss-5779be773682------2</link>
            <guid isPermaLink="false">https://medium.com/p/e56817d610ef</guid>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[chatgpt]]></category>
            <category><![CDATA[ai-and-education]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[writing]]></category>
            <dc:creator><![CDATA[Simon Hardy]]></dc:creator>
            <pubDate>Fri, 14 Jul 2023 14:09:55 GMT</pubDate>
            <atom:updated>2024-01-10T18:29:41.619Z</atom:updated>
            <content:encoded><![CDATA[<p><em>Unleash the power of AI without losing your human touch. Discover how to use ChatGPT as a collaborative tool, not a crutch.</em></p><h3>Introduction</h3><p>Today, I found myself in a fascinating conversation. I was lucky enough to have someone reading my articles and asking me the following: “Great writing throughput BTW — what’s your secret?” Around the same time, I stumbled upon an article published yesterday in the MIT Technology Review stating, “ChatGPT can turn bad writers into better ones. People who use ChatGPT to help with writing tasks are more productive and produce higher-quality work than those who don’t, a study found.”</p><p>This synchronicity got me thinking about the role of AI in our lives, particularly in writing.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*IJbGEuqWXpN2fDUmqYbGGw.png" /><figcaption>DALL-E Generated: “A piece of paper with abstract forms and binary numbers “by” <a href="https://en.wikipedia.org/wiki/Vera_Moln%C3%A1r">Vera Molnar</a>”</figcaption></figure><h3>1. AI: The Assistant, Not the Replacement</h3><p>Artificial Intelligence, particularly ChatGPT, has become an invaluable assistant. It can enhance our capabilities, including writing, by providing suggestions, correcting grammar, and even generating content. But it’s crucial to remember that AI is just that — an assistant. It can never replace the depth and breadth of human knowledge, the creativity, and the unique perspectives we bring to the table.</p><h3>2. The Double-Edged Sword of AI</h3><p>We’ve all heard the adage, “Those with power don’t want to lose it.” In our digital age, this can be transformed to “those with the technical means want to keep us captive in their environment.” This applies to AI as well.</p><p>When I write an article, I first come up with an idea and then mobilize my expertise, knowledge, and the advice I’ve received throughout my career. One piece of advice that has been instrumental in my writing journey and overall ability to convey my ideas was to read the book “Made to Stick” by the Heath Brothers.</p><h3>3. The Art of Writing: A Human Endeavor</h3><p>Writing is a deeply human endeavour. It involves the mechanical act of putting words on paper and the emotional journey of expressing our thoughts, ideas, and feelings. This is something that AI, for all its capabilities, cannot replicate.</p><p>When I use ChatGPT, I use it as a tool to review my article based on the methods from the best books and techniques for writing. I asked it to provide me with a score by dimensions and a global scoring to see what I could have done better. This is a way to keep learning and avoid creating a co-dependency with AI tools.</p><h3>4. Tailoring Content for the Audience: The Power of Prompts</h3><p>When writing, it’s crucial to consider the audience. Different platforms and publications have unique reader demographics and preferences. To ensure my content resonates with the intended audience, I’ve developed two prompts: one for my Medium posts and another for articles aiming to be published in recognized newspapers.</p><h4>A. Crafting Engaging Medium Posts</h4><p>Medium is a platform that encourages personal stories, insights, and unique perspectives. The audience here appreciates authentic, relatable content that provides value. Here’s the prompt I use to guide my writing for Medium:</p><ol><li>What personal story or experience can I share that relates to the topic?</li><li>How can I present my unique insights or perspective on this issue?</li><li>What actionable advice or takeaways can I provide to the readers?</li><li>How can I engage with the readers, encourage discussion, and build a community around this topic?</li></ol><p>See below an example of a prompt I crafted to score such articles:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/492/1*TX2cvrO4GHMRE332ecTadg.png" /><figcaption>Based on our discussion, I am more than flattered by this little “brainlike”.</figcaption></figure><h4>B. Writing for Recognized Newspapers</h4><p>Writing for recognized newspapers requires a different approach. The audience here expects well-researched, informative, and objective content. Here’s the prompt I use for these articles (the ones I am preparing behind the curtain for the moment):</p><ol><li>What is the key message or information I want to convey?</li><li>What research or data can I include to support my points?</li><li>How can I present this information concisely and objectively?</li><li>How can I engage the readers and encourage them to think more deeply about the issue?</li></ol><p>By tailoring my content to the audience, I can ensure that my articles resonate with the readers, provide value, and make an impact.</p><p>See below an example of a prompt I crafted to score such articles:</p><pre>Title of Your Article: _________<br><br>Link to Your Article: _________<br><br>Based on your input, please evaluate the suitability of the provided article for a high-quality publication such as The Economist or Politico, considering the following criteria inspired by noted resources:<br><br>Content Relevance and Depth (from &quot;On Writing Well: The Classic Guide to Writing Nonfiction&quot; by William Zinsser)<br><br>Score (1-10): _________<br>Feedback and suggestions: _________<br>Structure and Flow (from &quot;The Elements of Style&quot; by William Strunk Jr.)<br><br>Score (1-10): _________<br>Feedback and suggestions: _________<br>Language and Style (from &quot;The Sense of Style: The Thinking Person&#39;s Guide to Writing in the 21st Century&quot; by Steven Pinker)<br><br>Score (1-10): _________<br>Feedback and suggestions: _________<br>Data and Evidence (from &quot;The Data Journalism Handbook&quot; by Jonathan Gray, Lucy Chambers, Liliana Bounegru)<br><br>Score (1-10): _________<br>Feedback and suggestions: _________<br>Visual Aids and Graphics (from &quot;The Wall Street Journal Guide to Information Graphics: The Dos and Don&#39;ts of Presenting Data, Facts, and Figures&quot; by Dona M. Wong)<br><br>Score (1-10): _________<br>Feedback and suggestions: _________<br>Conclusions and Takeaways (from &quot;The Craft of Research&quot; by Wayne C. Booth, Gregory G. Colomb, and Joseph M. Williams)<br><br>Score (1-10): _________<br>Feedback and suggestions: _________<br>Overall Suitability Score (1-10): _________<br><br>Through this evaluation, aim to identify strengths and areas of improvement to refine the article to align with the standards of a professional publication.</pre><h3>5. The Need for Education and Collaboration: A Two-Way Street</h3><p>The key to harnessing the power of AI tools like ChatGPT lies in their use and understanding of their workings. Educating citizens on how to use these tools collaboratively is crucial, but the responsibility doesn’t end there. The AI tools should be designed to explain their processes to demystify how they improve or create text. This transparency fosters trust, encourages active involvement, and promotes a deeper understanding of the technology.</p><p>Imagine a world where AI tools don’t just correct your grammar or suggest better phrasing. They explain why a particular sentence structure works better or how using a specific word enhances the impact of your message. They become tools and teachers, imparting valuable knowledge about language and communication.</p><p>Moreover, if used correctly, AI tools can democratize education, breaking down barriers to access and affordability. They can serve as virtual tutors, providing personalized guidance and feedback, and mentors, offering insights and advice based on vast knowledge repositories.</p><p>However, this democratization of education should not be limited to a privileged few with technical know-how. It should be accessible to all, regardless of their background or expertise. This necessitates a radical rethinking of how we design and use AI tools. They should be intuitive, user-friendly, and inclusive, catering to diverse users.</p><p>In this way, we can ensure that the benefits of AI extend beyond productivity and efficiency. They become tools for lifelong learning, fostering a culture of continuous education and collaboration</p><h3>6. The Potential of AI: Democratizing Education</h3><p>ChatGPT and similar AI tools have the potential to democratize education, bringing it within reach of those who previously had limited access. However, they’re merely tools, not panaceas. This necessitates a radical rethinking of education — focusing on imparting knowledge and cultivating the ability to harness these tools effectively and ethically.</p><h3>Conclusion — The Future of AI: A Collaborative Approach</h3><p>We must remember that AI is not our adversary or replacement as we progress. When used correctly, it’s a tool that can enhance our capabilities and open up new possibilities. But it’s equally important to remember that we are not mere interfaces. We are human beings, capable of creativity, empathy, and innovation that no AI can replicate.</p><p>Interestingly, in a recent discussion with my psychologist, we pondered whether using tools like ChatGPT was helping me keep my perfectionism at bay or reinforcing it. It’s a thought-provoking question that underscores our complex relationship with AI. I’m sure I’ll delve into this topic in a future article.</p><h3>Resources</h3><ul><li>“<a href="https://www.technologyreview.com/2023/07/13/1076199/chatgpt-can-turn-bad-writers-into-better-ones/?truid=3d5aa974aa5b4709ae61822bb94286fe&amp;utm_source=the_download&amp;utm_medium=email&amp;utm_campaign=the_download.unpaid.engagement&amp;utm_term=Expire%20Suspends&amp;utm_content=07-14-2023&amp;mc_cid=1c808de88a&amp;mc_eid=4d96122094">ChatGPT can turn bad writers into better ones</a>”, MIT Technology Review, Rhiannon Williams, July 13th, 2023;</li><li>“<a href="https://www.academia.edu/5120191/Made_To_Stick_PDF">Made to stick: why some ideas survive and others die</a>”, Random House, Chip Heath and Dan Heath, 2007;</li><li>Nota Bene: you can find the other resources in the prompt template in this article.</li></ul><p><strong><em>Author Bio</em></strong><em>: I am the founder of </em><a href="https://www.citizeai.com/"><em>CitizeAi_</em></a><em>. My mission is to democratize AI and make it an accessible tool for everyone. My approach to AI and Data Strategy is rooted in humanistic principles, enabling me to develop solutions with a positive societal impact. I believe in transparency, self-awareness, and vulnerability and strive to foster open discussions on various topics. I invite you to connect with me as we shape the future of AI together. Please note that the views and opinions expressed in this article are those of </em><a href="https://www.citizeai.com/"><em>CitizeAi_</em></a><em> and my own and do not necessarily reflect the official policy or position of any other agency, organization, employer or company.</em></p><p><em>More content at </em><a href="https://plainenglish.io/"><strong><em>PlainEnglish.io</em></strong></a><em>.</em></p><p><em>Sign up for our </em><a href="http://newsletter.plainenglish.io/"><strong><em>free weekly newsletter</em></strong></a><em>. Follow us on </em><a href="https://twitter.com/inPlainEngHQ"><strong><em>Twitter</em></strong></a>, <a href="https://www.linkedin.com/company/inplainenglish/"><strong><em>LinkedIn</em></strong></a><em>, </em><a href="https://www.youtube.com/channel/UCtipWUghju290NWcn8jhyAw"><strong><em>YouTube</em></strong></a><em>, and </em><a href="https://discord.gg/GtDtUAvyhW"><strong><em>Discord</em></strong></a><strong><em>.</em></strong></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=e56817d610ef" width="1" height="1" alt=""><hr><p><a href="https://ai.plainenglish.io/the-power-and-peril-of-ai-harnessing-chatgpt-as-a-collaborative-tool-not-a-crutch-e56817d610ef">The Power and Peril of AI: Harnessing ChatGPT as a Collaborative Tool, Not a Crutch</a> was originally published in <a href="https://ai.plainenglish.io/">Artificial Intelligence in Plain English</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Invisible Citizens of the Digital Landscape: Bridging the AI Divide]]></title>
            <link>https://ai.plainenglish.io/the-invisible-citizens-of-the-digital-landscape-bridging-the-ai-divide-4009029bbbd6?source=rss-5779be773682------2</link>
            <guid isPermaLink="false">https://medium.com/p/4009029bbbd6</guid>
            <category><![CDATA[ai-governance]]></category>
            <category><![CDATA[digital-society]]></category>
            <category><![CDATA[data-ethics]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[inclusivity]]></category>
            <dc:creator><![CDATA[Simon Hardy]]></dc:creator>
            <pubDate>Thu, 13 Jul 2023 14:40:06 GMT</pubDate>
            <atom:updated>2024-01-10T18:31:08.266Z</atom:updated>
            <content:encoded><![CDATA[<p><em>Dive into the invisible side of AI. Ahmed’s story reveals the urgent call for inclusivity in our digitized world.</em></p><h4>Introduction</h4><p>In the heart of Stockholm, Sweden, Ahmed, a non-EU citizen, navigates his daily life in the captivating dance of the digital revolution. As the city pulses with the rhythm of Artificial Intelligence (AI) and data-driven decision-making, Ahmed finds himself in the shadows, an anomaly unregistered in the swirling data ocean. The prospect of Ahmed being completely overlooked by an AI-powered world is a chilling paradox we cannot ignore.</p><h4>1. Echoes in the Data Wilderness</h4><p>The Price of Being Unseen In Ahmed’s world, the data storm swirling around him carries a transformative power. Yet, paradoxically, he exists outside of its reach, a human story untold. His life goes uncounted in a world where data points determine identity, decisions, and rights.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*N0ZOMNKrvd19K36G7xNWoA.png" /><figcaption>DALL-E Generated: “A laptop is sending away a myriad of coloured numbers in waves “by” <a href="https://www.artnet.com/artists/julio-le-parc/2">Julio Le Parc</a>.”</figcaption></figure><h4>2. Shattering the Mirrors of ExclusivityThe Domino Effect</h4><p>Picture this: a digital world where AI systems merely reflect the faces of those recognized as ‘numbers’ in their quest for efficient management. Non-acknowledged citizens like Ahmed fall into a digital blind spot. This selective mirror, reflecting only parts of our societies, threatens to distort our collective reality, fragmenting it into ‘us’ and ‘them’.</p><p>The risks are chillingly real:</p><ul><li><strong>Discrimination</strong>: Ahmed’s children face unequal access to quality education and public services, their potential stunted before it can bloom;</li><li><strong>Invisibility</strong>: Ahmed’s needs and rights are overlooked, and his voice is drowned in the digital noise;</li><li><strong>Powerlessness</strong>: Ahmed is increasingly marginalized as AI systems become societal gatekeepers;</li><li><strong>Inequality</strong>: Ahmed’s lack of digital presence amplifies the global disparities, creating chasms in our increasingly interconnected world.</li></ul><p>What, then, is the role of AI in ensuring inclusivity, fairness, and social justice?</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*WlfbZMHL8BCOenoPFtFynA.png" /><figcaption>DALL-E Generated: “Bucket of binary data “by” <a href="https://www.artnet.com/artists/bridget-riley/">Bridget Riley</a>.”</figcaption></figure><h4>3. The Digital Compass: AI as a Tool for Inclusive Growth</h4><p>Could AI, currently a divisive force, be our tool for bridging divides? As an activist in Stockholm, Anna employs AI to fight against marginalization. She believes that AI development, grounded in critical thinking and equality, can prevent the hijacking of automated systems by extremist ideologies, making it a tool for unity.</p><h4>4. Redefining Power and Citizenship in the Age of AI</h4><p>The power dynamics shift in a world where AI systems hold sway over societal structures. An article in the MIT Technology Review paints AI not as a mechanical menace but as a political tool with the potential to alleviate democratic dissonance. Ahmed’s lack of ‘digital citizenship’ could be addressed through technical innovation and shifting societal power structures.</p><h4>Conclusion: Navigating the Digital Rapids Towards an Inclusive Future</h4><p><strong>Riding the Wave of Digital Revolution: </strong>How can we ensure Ahmed is included in this fast-paced AI revolution? How can we harness AI to pave the way for a future where no one is invisible?</p><p><strong>Unmasking Global Complexities</strong>: Ahmed may be recognized in Sweden in our interconnected digital world but still needs to be seen in the United States or China due to conflicting AI protocols. Navigating these complexities, we must strive for consistent and fair treatment of all individuals across global AI systems.</p><p><strong>Calling for Global Collaboration</strong>: An urgent call echoes in the digital wilderness for a standardized AI system, similar to ISO norms, to ensure seamless technology integration, akin to the unsung heroes of our time working behind the scenes.</p><p><strong>A Global Blueprint for AI</strong>: Together, we can draft a universal language for AI systems, guaranteeing fairness and inclusivity, not merely as technical requirements but as fundamental human rights.</p><p><strong>Harnessing the Potential of AI</strong>: Guided by this blueprint, we can leverage AI to redefine citizenship and bring about an equitable future.</p><p><strong>Unifying Threads of the Past and Future</strong>: Just as standardized nails bridged gaps during the Cold War, a global AI standard could unite us in our increasingly polarized world. Seizing this opportunity requires collective action. Together, we can ensure that the promise of the digital revolution includes Ahmed and the millions of others currently unseen in our data-driven world.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*HX5OC6u96ZQfUXFbQmIz6g.png" /><figcaption>DALL-E Generated: “A red nail on a world map with black and blue lines for the continents by <a href="https://www.artnet.com/artists/bridget-riley/">Bridget Riley</a>.”</figcaption></figure><p>The story of Ahmed, a single thread in our digital tapestry, reveals the challenges and potentials of AI. Let’s weave a future where every line is recognized, creating a vibrant, inclusive pattern in the age of AI. We have the loom, and we have the threads. Now, let’s create a masterpiece of digital inclusion.</p><h4>Resources</h4><ul><li><a href="https://link.springer.com/chapter/10.1007/978-3-030-41792-5_4">Data-Driven Government: The Triumph of Thatcherism or the Revenge of Society?</a>, Kieron O’Hara, 2020;</li><li><a href="https://academic.oup.com/edited-volume/37078">The Oxford Handbook of Digital Ethics</a>, Carissa Véliz, 2021;</li><li><a href="https://www.researchgate.net/publication/371870883_Social_AI_and_the_Challenges_of_the_Human-AI_Ecosystem">Social AI and the Challenges of the Human-AI Ecosystem</a>, Dino Pedreschi, Luca Pappalardo, Ricardo Baeza-Yates, Albert-Laszlo Barabasi, 2023;</li><li><a href="https://www.researchgate.net/publication/370887314_From_fear_to_action_AI_governance_and_opportunities_for_all">From fear to action: AI governance and opportunities for all</a>, Kevin Baum, Joanna J Bryson, Frank Dignum, Virginia Dignum, 2023;</li><li><a href="https://mitsloan.mit.edu/press/engineering-rules-a-new-book-mit-sloan-school-management-explores-how-often-overlooked-process-standardization-helps-accelerate-global-economic-growth">Engineering Rules: Global Standard Setting since 1880</a> (Hagley Library Studies in Business, Technology, and Politics), JoAnne Yates &amp; Craig Murphy, 2019;</li><li>“<a href="https://www.technologyreview.com/2021/04/23/1023549/kate-crawford-atlas-of-ai-review/">Stop talking about AI ethics. It’s time to talk about power</a>”, MIT Technology Review, Karen Hao, April 23, 2021.</li></ul><p><strong><em>Author Bio</em></strong><em>: I am the founder of </em><a href="https://www.citizeai.com/"><em>CitizeAi_</em></a><em>. My mission is to democratize AI and make it an accessible tool for everyone. My approach to AI and Data Strategy is rooted in humanistic principles, enabling me to develop solutions with a positive societal impact. I believe in transparency, self-awareness, and vulnerability and strive to foster open discussions on various topics. I invite you to connect with me as we shape the future of AI together. Please note that the views and opinions expressed in this article are those of </em><a href="https://www.citizeai.com/"><em>CitizeAi_</em></a><em> and my own and do not necessarily reflect the official policy or position of any other agency, organization, employer or company.</em></p><p><em>More content at </em><a href="https://plainenglish.io/"><strong><em>PlainEnglish.io</em></strong></a><em>.</em></p><p><em>Sign up for our </em><a href="http://newsletter.plainenglish.io/"><strong><em>free weekly newsletter</em></strong></a><em>. Follow us on </em><a href="https://twitter.com/inPlainEngHQ"><strong><em>Twitter</em></strong></a>, <a href="https://www.linkedin.com/company/inplainenglish/"><strong><em>LinkedIn</em></strong></a><em>, </em><a href="https://www.youtube.com/channel/UCtipWUghju290NWcn8jhyAw"><strong><em>YouTube</em></strong></a><em>, and </em><a href="https://discord.gg/GtDtUAvyhW"><strong><em>Discord</em></strong></a><strong><em>.</em></strong></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=4009029bbbd6" width="1" height="1" alt=""><hr><p><a href="https://ai.plainenglish.io/the-invisible-citizens-of-the-digital-landscape-bridging-the-ai-divide-4009029bbbd6">The Invisible Citizens of the Digital Landscape: Bridging the AI Divide</a> was originally published in <a href="https://ai.plainenglish.io/">Artificial Intelligence in Plain English</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[AI and the Welfare State: A New Paradigm for Universal Basic Income]]></title>
            <link>https://ai.plainenglish.io/ai-and-the-welfare-state-a-new-paradigm-for-universal-basic-income-b5dabfc42692?source=rss-5779be773682------2</link>
            <guid isPermaLink="false">https://medium.com/p/b5dabfc42692</guid>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[european-union]]></category>
            <category><![CDATA[democracy]]></category>
            <category><![CDATA[ai-and-ethics]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <dc:creator><![CDATA[Simon Hardy]]></dc:creator>
            <pubDate>Tue, 11 Jul 2023 12:00:49 GMT</pubDate>
            <atom:updated>2024-01-10T18:31:42.742Z</atom:updated>
            <content:encoded><![CDATA[<p><em>Unseen labour for AI: Time to rethink welfare and universal basic income in the digital age.</em></p><h4><strong>Introduction</strong></h4><p>Imagine being asked to work for free. Your immediate response would likely be a resounding “no.” Yet, in the digital age, we unknowingly contribute our labour to tech giants without compensation. This article explores how the rise of AI is reshaping our understanding of work, welfare, and the concept of a universal basic income.</p><h4><strong>1. The Unseen Workforce</strong></h4><p>Every time we solve a CAPTCHA, we are proving our humanity <strong>and contributing</strong> to the vast databases that fuel AI systems. These databases, in turn, are used to develop products and services we are asked to purchase. This form of labour goes unrecognized and uncompensated, new digital serfdom. But it’s not just about CAPTCHA. Whenever we use a search engine, interact with a virtual assistant or even browse through our social media feeds, we feed data into the AI systems that power these platforms. This data is then used to train and improve these systems, effectively making us unpaid workers in the AI economy.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*eZ5oaDLWFUqFmQMAlXJyLQ.png" /><figcaption>DALL-E Generated: “Waves of binary data by <a href="https://www.artnet.com/artists/jules-olitski/">Jules Olitski</a>.”</figcaption></figure><h4><strong>2. AI as a Societal Accelerator</strong></h4><p>If harnessed correctly, AI has the potential to bring about societal good and usher in a new era for the welfare state. The European Union, focusing on human rights and social welfare, has a crucial role in this transformation. The key is to ensure that those who contribute to the development of AI and its subsequent services/products receive fair compensation. This is not just about financial compensation but also about recognizing the value of the data we generate and our role in training AI systems. It’s about shifting the balance of power from the tech giants to the individuals who use their platforms.</p><h4><strong>3. From Universal Basic Income to Just Remuneration</strong></h4><p>The conversation around universal basic income needs to evolve. Instead of providing a basic income regardless of contribution, we should consider a system of just remuneration, where individuals are compensated for their contributions to AI development. This is not a handout but a fair exchange for work done. It’s about recognizing that our data and interactions with AI systems have value and that we should be compensated for this. It’s about moving away from a passive income to a more active form of remuneration, where our contributions to the AI economy are recognized and rewarded.</p><h4><strong>4. AI as a Guardian of Equality</strong></h4><p>AI can be a neutral actor promoting equality in our systems if programmed with the right values, such as critical thinking, European values, and Sustainable Development Goals. It can help enforce a more equitable distribution of resources and reduce growing inequalities, offering hope for future generations. But this is something that will only happen sometimes. It requires careful thought and planning and a commitment to ensuring that AI is used for the benefit of all, not just a privileged few.</p><h4><strong>5. The Role of the European Union and Its Citizens</strong></h4><p>The European Union, with its commitment to human rights and social welfare, has a crucial role to play in this transformation. It can set the standards for how AI is developed and used and ensure that the benefits of AI are shared fairly among all citizens. It can also provide a model for other regions and countries to follow, showing how AI can promote social good and not just corporate profits. However, this transformation cannot be achieved by the European Union alone. It requires the active participation and understanding of its citizens. This is where <strong>CitizeAi_</strong> comes into play, aiming to raise awareness about AI among the general public using simple, jargon-free language.</p><ul><li><strong>Public servants and experts</strong> also have a crucial role to play. They must provide clear and accessible explanations about AI and its implications, helping citizens understand how it affects their lives and society;</li><li><strong>The media</strong>, too, has a significant role to play. By providing accurate and understandable coverage of AI developments, they can help demystify the field and make it more accessible to the general public;</li><li><strong>Politicians</strong>, for their part, need to engage with the topic of AI positively and constructively. Instead of using AI to stoke fear and populism, they should focus on its potential to bring about societal good and promote a fairer distribution of resources.</li></ul><p>In this way, the development and use of AI can become a collective endeavour involving not just the European Union and tech companies but also citizens, public servants, experts, the media, and politicians. Together, they can ensure that AI is used in a way that benefits all and contributes to a more equitable society.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*erzEswwqS3qBY-Ze0calVw.png" /><figcaption>DALL-E Generated: “The European map by <a href="https://www.artnet.com/artists/bridget-riley/">Bridget Riley</a>.”</figcaption></figure><h4><strong>Conclusion: A Hope for the Future</strong></h4><p>We are living in a system with limited resources and growing inequalities. If understood and implemented correctly, AI can act as a guardian enforcing more equality in our systems; This is a hope for future generations. As we continue to develop AI, we must ensure that it serves us all equally and that its benefits are shared fairly.</p><h4><strong>Unveiling the AI Enigma: A Documentary in the Making</strong></h4><p>At CitizeAi_, We’re on a mission to demystify AI. Imagine a world where every citizen, regardless of background, can grasp the intricacies of Artificial Intelligence. In this world, AI is not a distant, complex concept but a part of our daily conversations. We’re making this vision a reality through an upcoming documentary.</p><p>This isn’t just a film; it’s a bridge connecting citizens to the realm of AI. It’s a journey of understanding designed to equip us with the knowledge to navigate the AI landscape. Stay tuned for a sneak peek into the world of AI like never before. Let’s unravel the AI mystery together.</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fplayer.vimeo.com%2Fvideo%2F849900006%3Fapp_id%3D122963&amp;dntp=1&amp;display_name=Vimeo&amp;url=https%3A%2F%2Fvimeo.com%2F849900006&amp;image=https%3A%2F%2Fi.vimeocdn.com%2Fvideo%2F1704110249-7c7e7820a6c0bd81f2dc67bd46cdda20859632d06fe45934f8b805ad35960a90-d_1280&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=vimeo" width="1920" height="1080" frameborder="0" scrolling="no"><a href="https://medium.com/media/0bdae30a55e5938b4decd13c80321a52/href">https://medium.com/media/0bdae30a55e5938b4decd13c80321a52/href</a></iframe><p>Stay curious, and stay informed. The future of AI is in our hands.</p><h4><strong>Resources:</strong></h4><ul><li>Watson, N., &amp; Bianca, D. (2018). <a href="http://arxiv.org/abs/1803.11258v1">Welfare Without Taxation — Autonomous production revenues for Universal Basic Income</a>;</li><li>Rasoolinejad, M. (2019). <a href="http://arxiv.org/abs/1910.05658v2">Universal Basic Income: The Last Bullet in the Darkness</a>;</li><li>Abebe, R., &amp; Goldner, K. (2018). <a href="http://arxiv.org/abs/1810.09832v1">Mechanism Design for Social Good</a>;</li><li>Tahaei, M., Constantinides, M., Quercia, D., &amp; Muller, M. (2023). <a href="http://arxiv.org/abs/2302.05284v3">A Systematic Literature Review of Human-Centered, Ethical, and Responsible AI</a>;</li><li>Malkov, E. (2021). <a href="http://arxiv.org/abs/2108.09981v2">Welfare Effects of Labor Income Tax Changes on Married Couples: A Sufficient Statistics Approach</a>.</li></ul><p><strong><em>Author Bio</em></strong><em>: I am the founder of </em><a href="https://www.citizeai.com/"><em>CitizeAi_</em></a><em>. My mission is to democratize AI and make it an accessible tool for everyone. My approach to AI and Data Strategy is rooted in humanistic principles, enabling me to develop solutions with a positive societal impact. I believe in transparency, self-awareness, and vulnerability and strive to foster open discussions on various topics. I invite you to connect with me as we shape the future of AI together. Please note that the views and opinions expressed in this article are those of </em><a href="https://www.citizeai.com/"><em>CitizeAi_</em></a><em> and my own and do not necessarily reflect the official policy or position of any other agency, organization, employer or company.</em></p><p><em>More content at </em><a href="https://plainenglish.io/"><strong><em>PlainEnglish.io</em></strong></a><em>.</em></p><p><em>Sign up for our </em><a href="http://newsletter.plainenglish.io/"><strong><em>free weekly newsletter</em></strong></a><em>. Follow us on </em><a href="https://twitter.com/inPlainEngHQ"><strong><em>Twitter</em></strong></a>, <a href="https://www.linkedin.com/company/inplainenglish/"><strong><em>LinkedIn</em></strong></a><em>, </em><a href="https://www.youtube.com/channel/UCtipWUghju290NWcn8jhyAw"><strong><em>YouTube</em></strong></a><em>, and </em><a href="https://discord.gg/GtDtUAvyhW"><strong><em>Discord</em></strong></a><strong><em>.</em></strong></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=b5dabfc42692" width="1" height="1" alt=""><hr><p><a href="https://ai.plainenglish.io/ai-and-the-welfare-state-a-new-paradigm-for-universal-basic-income-b5dabfc42692">AI and the Welfare State: A New Paradigm for Universal Basic Income</a> was originally published in <a href="https://ai.plainenglish.io/">Artificial Intelligence in Plain English</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Ethically Aligned Pyramid Builder: A New Era of AI Governance and Data Spaces]]></title>
            <link>https://medium.com/@mesimonhardy/the-ethically-aligned-pyramid-builder-a-new-era-of-ai-governance-and-data-spaces-c8ca578d15c7?source=rss-5779be773682------2</link>
            <guid isPermaLink="false">https://medium.com/p/c8ca578d15c7</guid>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[dataspace]]></category>
            <category><![CDATA[ai-and-ethics]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[european-union]]></category>
            <dc:creator><![CDATA[Simon Hardy]]></dc:creator>
            <pubDate>Sun, 02 Jul 2023 17:18:47 GMT</pubDate>
            <atom:updated>2024-01-10T18:32:14.954Z</atom:updated>
            <content:encoded><![CDATA[<p><em>Discover the Ethically Aligned Pyramid Builder, a new era of AI governance and data spaces where ethics and technology merge.</em></p><h3>Introduction</h3><p>In AI, Data Strategy, and Governance, I often draw parallels with an intriguing historical counterpart — the pyramid builder. This comparison, while seemingly unlikely, is not entirely far-fetched. A genetic report from 23andMe once revealed my surprising lineage back to ancient Egypt. I imagine my ancestors would find it fascinating how their craft has evolved into shaping the digital landscapes of today.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/707/1*4UU_zMEMWLT8MeFpfEMD4A.gif" /><figcaption>What should you see at the top?</figcaption></figure><h3>1. The Modern Pyramid: A Triad of Ethical Clusters</h3><p>In our modern pyramid, the first stone isn’t a physical block but a triad of ethical clusters that guide our AI systems: Rational Thinking, European Values, and the United Nations’ Sustainable Development Goals (SDGs). These clusters are not hierarchical but coexist at the top of the pyramid, each equally important in guiding the ethical alignment of our AI systems.</p><h3>2. Rational Thinking: The Bedrock of AI Systems</h3><p>Inspired by works like Popper’s “The Logic of Scientific Discovery” and Kahneman’s “Thinking, Fast and Slow,” the Rational Thinking cluster comprises principles like transparency, empirical validation, and expert knowledge. We score each algorithm based on these principles from 1 to 3 (with 1 being weak, 2 being moderate, and 3 being strong), a mechanism to gauge how well it’s grounded in rational thought.</p><h3>3. European Values: The Ethical Compass of AI Systems</h3><p>The European Values cluster, referencing Dahl’s “Democracy and its Critics” and the EU’s Charter of Fundamental Rights, stands for respect for user autonomy, solidarity, and justice. Each algorithm is assessed based on these principles, providing another score. Here, we measure the alignment of the algorithm with the ethical and societal standards intrinsic to the EU.</p><h3>4. United Nations’ SDGs: The Global Standard for AI Systems</h3><p>The United Nations’ SDGs cluster symbolizes global ethical standards, with sustainability and inclusiveness at its core. As before, we evaluate and score the algorithm on these principles, ensuring the algorithm aligns with international standards for ethical development.</p><h3>5. The All-Seeing Eye: The Overall Score</h3><p>Just as the all-seeing eye atop the Great Seal of the United States pyramid symbolizes providence and guidance, our overall score offers an immediate understanding of an algorithm’s ethical alignment. The final score, derived from a specific calculation, succinctly encapsulates the ethical standing of the algorithm.</p><h3>6. Embedding Ethical Guardrails in AI Systems</h3><p>These ethical clusters can be embedded as guardrails in any AI system. Before implementing algorithms, these prompts ensure that our AI systems are ethically aligned and adhere to the principles of rational thought, European values, and the United Nations’ SDGs.</p><pre>Algorithm Name: _________<br><br>Algorithm Description: _________<br><br>Please assess the algorithm based on the following clusters of principles:<br><br>Cluster 1: Rational Thinking<br><br>Transparency (inspired by Popper&#39;s &quot;The Logic of Scientific Discovery&quot;) <br>- Score (1-3):<br><br>Explanation: <br>Bias and Fairness <br>(inspired by Kahneman&#39;s &quot;Thinking, Fast and Slow&quot; and <br>Yudkowsky&#39;s &quot;Rationality: From AI to Zombies&quot;) - Score (1-3):<br><br>Explanation: ________<br>Empirical Validation (inspired by Sagan&#39;s <br>&quot;The Demon-Haunted World: Science as a Candle in the Dark&quot; <br>and &quot;The Book of Why: The New Science of Cause and Effect&quot; by Judea Pearl <br>and Dana Mackenzie) - Score (1-3):<br><br>Explanation: _________<br>Use of Expert Knowledge (inspired by Nichols&#39; <br>&quot;The Death of Expertise: The Campaign Against Established Knowledge and <br>Why it Matters&quot;) - Score (1-3):<br><br>Explanation: _________ <br><br>Cluster 2: European Values<br><br>Respect for User Autonomy (inspired by Dahl&#39;s <br>&quot;Democracy and its Critics&quot;, <br>Haidt&#39;s &quot;The Righteous Mind: Why Good People Are Divided by Politics <br>and Religion&quot;, and &quot;Automating Inequality&quot; by Virginia Eubanks) - <br>Score (1-3):<br><br>Explanation: _________<br>Solidarity (inspired by EU&#39;s Charter of Fundamental Rights) - Score (1-3):<br><br>Explanation: _________<br>Justice (inspired by EU&#39;s Charter of Fundamental Rights <br>and &quot;Fairness and Machine Learning: Limitations and Opportunities&quot; <br>by Solon Barocas, Moritz Hardt, and Arvind Narayanan) - Score (1-3):<br><br>Explanation: _________<br><br>Cluster 3: United Nations&#39; SDGs<br><br>Sustainability (inspired by United Nations&#39; SDGs) - Score (1-3):<br><br>Explanation: _________<br>Inclusiveness (inspired by United Nations&#39; SDGs) - Score (1-3):<br><br>Explanation: _________<br>Overall Score (1-10): _________ <br>(sum of scores divided by 9, then multiplied by 2/3)<br><br>A brief summary of the overall alignment of the algorithm with the <br>principles outlined: _________<br><br>This structure allows for a clear understanding of how well the <br>algorithm aligns with each set of principles and provides a comprehensive <br>assessment of its overall alignment.</pre><h3>7. Anchoring Ethical Logic in Data Spaces</h3><p>This kind of logic must be anchored in the Data Spaces to ensure adequate usage of data. Like a vast library, the European Data Spaces (EDS) provide a framework for secure and standardized data exchange across different sectors. By embedding our ethical clusters in the EDS, we can ensure that our AI systems are ethically aligned and make the best use of available data.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/333/1*I3n1-ERBPPaNGttTKUGj0w.png" /><figcaption>First things first.</figcaption></figure><h3>Conclusion</h3><p>As pyramid builders of the digital age, we understand that our efforts today will dictate the landscapes of tomorrow. Using our tools and principles, we construct edifices that withstand the test of time and uphold the values we wish to propagate in this AI-driven era. Our creations, like the pyramids of my distant ancestors, will stand tall, embodying the values and visions we want to carry forward into the digital future.</p><h3>Resources</h3><ul><li>Popper, K. (1959). “<a href="https://en.wikipedia.org/wiki/The_Logic_of_Scientific_Discovery">The Logic of Scientific Discovery.</a>” Hutchinson &amp;Co;</li><li>Kahneman, D. (2011). “<a href="https://en.wikipedia.org/wiki/Thinking,_Fast_and_Slow">Thinking, Fast and Slow.</a>” Farrar, Straus and Giroux;</li><li>Dahl, R. (1989). “<a href="https://en.wikipedia.org/wiki/Democracy_and_Its_Critics">Democracy and its Critics</a>.” Yale University Press;</li><li><a href="https://www.europarl.europa.eu/charter/pdf/text_en.pdf">Charter of Fundamental Rights of the European Union</a>;</li><li><a href="https://www.undp.org/sustainable-development-goals">United Nations’ Sustainable Development Goals</a>. United Nations.</li></ul><p><strong><em>Author Bio</em></strong><em>: I am the founder of </em><a href="https://www.citizeai.com/"><em>CitizeAi_</em></a><em>. My mission is to democratize AI and make it an accessible tool for everyone. My approach to AI and Data Strategy is rooted in humanistic principles, enabling me to develop solutions with a positive societal impact. I believe in transparency, self-awareness, and vulnerability and strive to foster open discussions on various topics. I invite you to connect with me as we shape the future of AI together. Please note that the views and opinions expressed in this article are those of </em><a href="https://www.citizeai.com/"><em>CitizeAi_</em></a><em> and my own and do not necessarily reflect the official policy or position of any other agency, organization, employer or company.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c8ca578d15c7" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Evolution of European Integration: From Judicialization to Data Spaces]]></title>
            <link>https://medium.com/@mesimonhardy/the-evolution-of-european-integration-from-judicialization-to-data-spaces-79d68c27873e?source=rss-5779be773682------2</link>
            <guid isPermaLink="false">https://medium.com/p/79d68c27873e</guid>
            <category><![CDATA[data]]></category>
            <category><![CDATA[democracy]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[european-union]]></category>
            <category><![CDATA[ai]]></category>
            <dc:creator><![CDATA[Simon Hardy]]></dc:creator>
            <pubDate>Fri, 30 Jun 2023 12:32:34 GMT</pubDate>
            <atom:updated>2024-01-10T18:32:40.220Z</atom:updated>
            <content:encoded><![CDATA[<p><em>Unravel the EU’s journey from Judicialization to Data Spaces, and discover how data is reshaping our democratic essence.</em></p><h3>Introduction</h3><p>I remember conversing with an Uber driver while working on the Public Procurement Data Space project. He complained about the EU, expressing his frustration about what he perceived as inefficiency and lack of transparency. As we navigated through the city, I found myself navigating a different terrain — the landscape of public perception of the EU. I took this opportunity to explain my work on the Public Procurement Data Space and how it aimed to protect public money — our money.</p><p>I told him to imagine the EU as a complex symphony of diverse nations, each playing its part harmoniously with the others. Over the years, the EU has evolved, adapting to the changing times and challenges. Two key phases in this evolution are the Judicialization of the EU in the 1970s and the recent development of the European Data Spaces (EDS). These two phases, although different, share a common goal: to strengthen the democratic essence of the EU and promote harmonized decision-making.</p><p>By the end of the ride, the driver’s perspective had shifted. He began to see the value in the work being done by the Commission, especially regarding protecting taxpayer money. This experience underscored the importance of making our work more accessible to the public. We must leave our echo chambers and engage in meaningful dialogues with citizens. After all, protecting taxpayer money is a cause that resonates with everyone.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*JEZ_MHygdKdUr6VxNhEgDg.gif" /><figcaption>1. 2.. 3… Open your mouth.</figcaption></figure><h3>1. The Judicialization of the EU — The Unseen Hand Guiding the EU Towards Greater Integration</h3><p>The 1970s was a pivotal decade for the EU. Amidst a political vacuum, a phenomenon known as the ‘Judicialization of the EU’ emerged, acting as the unseen hand guiding the EU towards greater integration. This process can be likened to a ship’s rudder, subtly yet effectively steering the EU in the desired direction. The Judicialization of the EU refers to the increasing role of judicial institutions, particularly the European Court of Justice (ECJ), in shaping EU policies and governance. During the 1970s, the ECJ interpreted European Community law in ways that expanded the powers of European institutions. This was akin to a sculptor carefully chiselling away to reveal the form within the stone — the form, in this case, being a more integrated and unified EU.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/500/1*yLMB81vIYmeHQi3sf30LRA.gif" /><figcaption>Keeping a straight head.</figcaption></figure><h4>The Power of Judicialization — The Case of Poland’s Rule of Law Dispute</h4><p>A striking example of the power of Judicialization is the recent case of Poland’s rule of law dispute. Imagine a game of chess where one player decides to change the rules midway. This is akin to what Poland did by disregarding the primacy of EU law and undermining judicial independence. The ECJ, acting as the conductor of the EU orchestra, decided to impose a fine on Poland, demanding that Poland fall back in line with the rest of the orchestra. The fine was intended to prevent “serious and irreparable harm” to the EU’s legal order and values, much like a conductor would strive to prevent a discordant note from ruining a symphony. The imposition of the fine demonstrates the power of the Judicialization of the EU, showing that the ECJ can exert significant influence over member states, ensuring they adhere to the EU’s democratic principles and the rule of law.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/229/1*Cai5t6bZ695i1Mz0AoIGAw.gif" /><figcaption>Chess, Law and Data.</figcaption></figure><h3>2. The European Data Spaces — A New Paradigm for Harmonized Decision-Making</h3><p>Fast forward to the present day, and the EU is on the cusp of another transformative phase — the creation of the European Data Spaces. The EDS is like a vast library, where every piece of data is a book that can provide valuable insights for decision-making. By providing a framework for secure and standardized data exchange across different sectors, the EDS ensures that decision-makers have access to the most accurate and reliable data, much like a librarian ensuring that readers have access to the best books.</p><p>The Role of Public Procurement Data Space in Identifying Fraud: One of the key sectors that can greatly benefit from the EDS is public procurement. Imagine a detective trying to solve a case. The detective needs clues, evidence, and information to piece together the puzzle and solve the issue. The Public Procurement Data Space is like the detective’s toolkit, providing all the data and tools to identify fraud, collusion, and misuse of European money.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/564/1*kTg-1Wo0vXP4c62C-SzAxg.gif" /><figcaption>A new era for the Rosetta Stone.</figcaption></figure><h4>The Case of Portugal</h4><p>A Real-World Example: Let’s take the case of Portugal as an example. The European Public Prosecutor’s Office (EPPO), akin to a vigilant watchman, has opened nine investigations in Portugal as of the end of 2021. These cases represent an estimated total damage of €158.2 million, a staggering amount that could have been used for the betterment of society. The crimes involved in these cases include fraud, VAT tax evasion, and money laundering, all complex and sophisticated crimes that require a comprehensive and effective investigation.</p><p>With the Public Procurement Data Space, the EPPO can leverage data to identify patterns, anomalies, and red flags that may indicate fraudulent activities. It’s like having a magnifying glass reveal hidden details and clues that might go unnoticed. The EPPO can detect irregularities, track money flow, and uncover potential fraud and corruption cases by analysing procurement data.</p><h3>Conclusion</h3><p>The evolution of the EU from Judicialization in the 1970s to the creation of the European Data Spaces is a testament to the EU’s adaptability and commitment to democratic principles. Like the Judicialization before it, the EDS is set to play a crucial role in guiding the EU towards greater integration and harmonized decision-making. As we move into the data and AI paradigm, the EDS will ensure that the democratic essence of the EU is embedded in every decision, preserving the harmony of the EU symphony. Looking into the future, we can imagine an even more automated decision-making process, where companies not respecting the Green Deal objectives could be automatically excluded, leading to a more comprehensive society and reducing democratic dissonance. We are building this future where the EU is not just a political entity but a data-driven, democratic, and inclusive community.</p><h3>Resources</h3><ul><li>Otto, B., ten Hompel, M., Wrobel, S. (eds) <a href="https://link.springer.com/book/10.1007/978-3-030-93975-5">Designing Data Spaces</a>. Springer, Cham.s;</li><li>Alter, K. J. (2009). <a href="https://global.oup.com/academic/product/the-european-courts-political-power-9780199558353?cc=be&amp;lang=en&amp;">The European Court’s Political Power</a>;</li><li>Stone Sweet, A. (2004). <a href="https://academic.oup.com/book/36192">The Judicial Construction of Europe</a>;</li><li>European Public Prosecutor’s Office (2022). <a href="https://eppo.europa.eu/en">Annual Report</a>;</li><li>European Commission, <a href="https://digital-strategy.ec.europa.eu/en/policies/strategy-data">European Data Spaces</a>.</li></ul><p><strong><em>Author Bio</em></strong><em>: I am the founder of </em><a href="https://www.citizeai.com/"><em>CitizeAi_</em></a><em>. My mission is to democratize AI and make it an accessible tool for everyone. My approach to AI and Data Strategy is rooted in humanistic principles, enabling me to develop solutions with a positive societal impact. I believe in transparency, self-awareness, and vulnerability and strive to foster open discussions on various topics. I invite you to connect with me as we shape the future of AI together. Please note that the views and opinions expressed in this article are those of </em><a href="https://www.citizeai.com/"><em>CitizeAi_</em></a><em> and my own and do not necessarily reflect the official policy or position of any other agency, organization, employer or company.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=79d68c27873e" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>