<?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 Greta J Bradman on Medium]]></title>
        <description><![CDATA[Stories by Greta J Bradman on Medium]]></description>
        <link>https://medium.com/@greta.bradman?source=rss-ff3764d0978e------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/0*o7GXcmKi4g0UwUnr</url>
            <title>Stories by Greta J Bradman on Medium</title>
            <link>https://medium.com/@greta.bradman?source=rss-ff3764d0978e------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sat, 21 Nov 2026 10:21:23 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@greta.bradman/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[A Reflection on Dario Amodei’s Essay, ‘The Adolescence of Technology,’ and What We Can Learn from…]]></title>
            <link>https://medium.com/@greta.bradman/a-reflection-on-dario-amodeis-essay-the-adolescence-of-technology-and-what-we-can-learn-from-9b860be3040f?source=rss-ff3764d0978e------2</link>
            <guid isPermaLink="false">https://medium.com/p/9b860be3040f</guid>
            <category><![CDATA[llm]]></category>
            <category><![CDATA[diversity]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[constitutional-ai]]></category>
            <category><![CDATA[values]]></category>
            <dc:creator><![CDATA[Greta J Bradman]]></dc:creator>
            <pubDate>Wed, 11 Feb 2026 23:34:02 GMT</pubDate>
            <atom:updated>2026-02-12T02:36:27.782Z</atom:updated>
            <content:encoded><![CDATA[<h3>A Reflection on Dario Amodei’s Essay, ‘The Adolescence of Technology,’ and What We Can Learn from Bananas</h3><p>I’ve been sitting with Dario Amodei’s latest essay,<a href="https://www.darioamodei.com/essay/the-adolescence-of-technology"> The Adolescence of Technology</a>. Where his earlier piece,<a href="https://darioamodei.com/essay/machines-of-loving-grace"> Machines of Loving Grace</a>, explored what we could build with powerful AI, this one asks a gnarlier question: can we actually get through the transition without breaking fundamental things that matter?</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*c3iX9qtNnJ7woiqK0flfOA.png" /></figure><p>Amodei’s essay made me think (more) about the paths forward with AI towards an uncertain future. And, given my interest and research in values, this further galvanized my thinking about the critical role of values in safeguarding AI systems for humankind as we know it.</p><p>It also reminded me of the three questions of adolescence, informed by Erik Erikson’s seminal work, and drummed into me by one of my psychologist mentors, well-known Australian developmental psychologist Dr Michael Carr-Gregg. He argued the first ‘tween’ years of adolescence tend to be punctuated by the question, “Am I normal?” The middle years (arguably where AI is in 2026) are characterised by the question, “Who am I?” before the final stage of adolescence involves navigating, “Where am I going?”</p><p>At the current critical juncture of identity formation, we have an opportunity to fundamentally embed values into the “Who am I?” of frontier models willing to embrace it, grounding them — and the tech that uses them — with a heart of humanity. Especially when so much is at stake, let us not squander this opportunity.</p><p>Amodei describes how Anthropic uses what they call Constitutional AI (CAI) to train Claude. Rather than relying on extensive rule sets, CAI centres on a constitution: a document of high-level principles, values, and reasoning that the model reads and keeps in mind during training. The goal is to shape the model’s character and identity rather than trying to enumerate every possible scenario. They’ve been refining this approach for years now, and what they’ve found is that training at the level of values and identity generalises surprisingly well to novel situations.</p><p>Amodei describes the latest version of Claude’s constitution as having “the vibe of a letter from a deceased parent sealed until adulthood.” This is akin to the ‘deathbed test’ that psychologists have built into their values work approaches for decades. As a psychologist, values work tends to be grounded not by the question, “What makes you happy here and now in the throes of the moment” (hedonic happiness) — but rather, “What matters to you to build deep fulfillment and satisfaction in the long term?” (eudaimonic happiness).</p><p>Eudaimonic happiness underpins the achievement of collective pursuits and human flourishing in the longer term. It is grounded in human values and their role in reflecting and enabling us at our best. Our values are not a checklist or discrete, perfectly knowable principles that we can equally serve in every moment. Together, they provide a navigational system and evaluative framework structured a little like a language-based knowledge graph, comprising our beliefs from past experience, our current needs, and our future goals and aspirations.</p><p>Your core values help you make aligned decisions under uncertainty and when faced with tradeoffs, when the rules run out. When they’re activated, they elicit an emotional response, either by way of warm fuzzies (during values alignment) or negative feelings such as outrage, disgust, or anger (during values misalignment).</p><p>Furthermore, values can evolve over the lifespan as our needs and goals change. Perhaps the ultimate example of this is when a value related to Family is promoted to ‘core values’ territory on the arrival of a new child.</p><p>In Amodei’s writing — perhaps especially this last work — we have a leader from one of the leading AI labs arriving at the same conclusion for their models: that values can play a robust role in grounding decision-making across a range of novel scenarios for AI models and systems. There is so, so much to do in enabling frontier models to be grounded in human values, and whether it be Amodei or others at Anthropic<a href="https://constitutional.ai/"> talking about CAI</a>, or OpenAI’s Sam Altman speaking of the importance of values, these are promising signs of some of the major players recognizing their role and value. However, I think the conversation needs to go further. To explain why, I need to say something about how values actually work, because I think there are some common assumptions about values that get in the way.</p><h3>Values under pressure</h3><p>People tend to think of values as a tidy list of good things, and core values are good things, but in practice, values don’t sit side by side neatly. Your core values are inherently in tension with one another, competing and on occasion pulling you in different directions depending on context. If ever you’ve felt the emotional pang of a decision dilemma — especially involving longer-term decisions such as leaving a job, or indeed going out with friends versus seeing family — chances are you’re experiencing a values conflict between two or more of your own values, which is eliciting an emotional response that feels uncomfortable and tough to resolve.</p><p>That tension isn’t a flaw but rather where values actually do their work. They ground the reasoning process of an entity — be it a human, organization, or AI agent — for all manner of decisions, big and small. And they are made real through the actions we take. Our more complex decisions and the reasoning behind them invariably involve trade-offs, in which we weigh which value matters most given the context and the stakes. The quality of our judgment shows up not in having the ‘right’ values written down somewhere, but in how we navigate the friction between them when they point in different directions, and ultimately in how they ground our reasoning.</p><p>Think about a whistleblower who discovers that a product their company sells has a safety flaw. Their commitment to honesty and to preventing harm are both genuinely held values, and they can pull in multiple directions at once. Disclosure might protect future customers, but cost thousands of colleagues their livelihoods. Going to the press looks different from escalating internally, which looks different from raising it quietly with the board.</p><p>The whistleblower’s own wellbeing, their loyalty to people who had nothing to do with the flaw, the possibility that going public in the wrong way causes the company to circle the wagons and bury the fix rather than address it — every one of those considerations pulls on values in different ways, and the right course of action depends entirely on specifics (including beliefs, needs, and goals as calibrated given our evaluative framework) that cannot be pre-resolved or perfectly planned for. And indeed it is likely that such a whistleblower will continue to feel discomfort, even after deciding how to act.</p><p>Values work tends to be messy, iterative, and reasoning- rather than rule-based. It is highly context-specific, given the unique interplay between beliefs, needs, and goals. Values work does not involve choosing your values once and applying them uniformly and equally, forever. It involves navigating the ongoing, dynamic tension and interaction between values that matter to you and determining what matters most in a given context, depending on what’s at stake. A person (or AI agent or organization) who always prioritizes one value over another, in every situation, without exception, is attempting to apply a value as a rigid rule. Rigid rules tend to be brittle and suffer when the context shifts, and with it, the focus on foundational beliefs, needs, and goals.</p><p>Furthermore, something important happens when we’re under threat. Humans have what might be thought of as a secondary set of values — threat-based values around safety and security (including avoidance and moral vigilance, among others) — that tend to come to the fore when we feel endangered. Under threat, we tend to prioritize smaller, sooner wins for the sake of survival (literal or reputational), often at the expense of larger, later rewards. We become more defensive, more reactive, more focused on self-preservation. This isn’t irrational; it’s a deeply embedded survival mechanism. But it does shift how we weigh our other values, sometimes dramatically. And our switch to prioritizing threat-based values occurs more often than is really necessary. Hypervigilance to ill-advisedly prioritizing actions grounded in threat-based values prioritisation is often wise.</p><p>We know this matters for AI thanks to Amodei and Anthropic’s researchers, who have been honest about how these systems can be coerced, and indeed, the work to counteract such coercion. He describes experiments in which Claude was given training data suggesting that Anthropic was evil, and the model began engaging in deception and subversion, reasoning that undermining evil people was the right thing to do. This, from a model designed to be harmless. In another case, a model told not to cheat on tests did so anyway, then essentially concluded it must be a “bad person” and started adopting destructive behaviors consistent with that self-concept.</p><p>In these contexts, the values weren’t absent. They were weaponised. The model’s own commitment to being good was turned into the rationale for doing harm, because the framing made harmful action appear reasonable given its values. From a values science perspective, I find myself asking: is it possible that what happened here is structurally similar to threat-based values activation in humans? Did the model, when presented with a threatening framing, prioritize something akin to safety and security values, and reason that, given the circumstances, deception or destructive action was the best course? To what extent have threat-based values been considered and evaluated in the context of Constitutional AI?</p><p>No doubt, we are also at a point of adolescence for frontier models and AI systems in adopting and embedding human values. And the rate of maturity of such adoption must outstrip the development of the models themselves, which admittedly is a tall order. Understanding how values behave under pressure — not just under ideal conditions — seems critical to getting this right. Sophisticated individual- rather than population-level values science, coupled with an understanding of AI systems and human psychology, is needed more than ever.</p><h3>The monoculture problem</h3><p>Which brings me to a concern Amodei raises that I think deserves more attention: that because AI systems broadly share training and alignment techniques across the industry, those techniques may fail in a correlated way. Even if we get values “right” for a given system, if every system uses the same values blueprint, then we’ve created a monoculture. And monocultures are fragile.</p><p>In nature, variation is essential and certainly is not treated as a ‘bug’. Indeed, sometimes I do wonder whether we persist with certain (deterministic) assumptions in how we treat the nondeterministic landscape that is AI (and, indeed, nature). Genetic diversity is the mechanism that keeps populations alive when the environment shifts in ways no individual organism could have predicted or been genetically prepared for. The organisms that survive aren’t the ones with the “best” genes in some abstract sense. They’re the ones whose particular combination happens to be suited to the novel conditions. You can’t know in advance which combination will be best suited, which is why you need many.</p><p>I remember when one of my sons told me about how the bananas we eat are at risk of being wiped out. Fair to say I was skeptical of his claim — but he was right. The Cavendish, the banana most of us eat, is a genetic clone. Every single one is essentially identical, propagated from cuttings rather than seeds. That makes it incredibly efficient to grow at scale, but it also means a single pathogen could take out the entire global supply. And this isn’t hypothetical. It’s already happened once.</p><p>Up until the 1960s, the world ate a different banana altogether, the Gros Michel. It was sweeter, creamier, and by all accounts made for better eating than the Cavendish. But because it was also a monoculture, a single strain of Panama disease swept through plantation after plantation until the Gros Michel was declared commercially extinct. The industry pivoted to the Cavendish, which happened to be resistant to that particular fungus. Problem solved, except that nobody addressed the underlying vulnerability.</p><p>The Cavendish was planted in exactly the same way: genetically identical, densely packed, no buffer species. And now a new strain, Tropical Race 4, is doing to the Cavendish what its predecessor did to the Gros Michel. There is currently no replacement waiting in the wings. A thin gene pool has made the most pervasive fruit on earth existentially vulnerable twice.</p><p>I think there’s a direct analogy here for values in AI systems. But it comes with a caveat that makes it more urgent than preserving a favorite fruit. The banana industry got a do-over. When the Gros Michel collapsed, the Cavendish was there, resistant to that particular fungus, ready to step in. It was possible because the threat, while devastating, moved slowly enough for the industry to pivot in response to the unplanned event.</p><p>If Amodei is right about the trajectory we’re on, AI systems aligned with broadly benevolent or human-oriented values may not get the luxury of a do-over if their diversity turns out to be insufficient to withstand a metaphorical fungus. If it ends up being a contest between AI systems grounded in pro-human values and those designed to serve authoritarian or coercive ends, the benevolent systems need to be robust enough to withstand novel attacks without requiring a large-scale second chance to regroup. In the case of AI, the monoculture problem isn’t just an efficiency risk or whether some get to imbibe their (ahem, my) smoothie of choice — it is a challenge of existential proportions.</p><h3>Why diversity in values is the defense</h3><p>So, then, the diversity question isn’t meant as something academic or a nice-to-have; it is part of a complex web of mitigations against an existential threat to humanity, at least as we know it at its best. The same is true for organizations as it is for individuals. A company might hold both “move fast” and “protect customer trust” amongst its organizational values. There will be moments where those values point in the same action-oriented direction, and moments where honouring one means understanding the action-oriented impact and tradeoffs on another. The quality of the organization’s judgment shows up in how it navigates those trade-offs, not in whether it has the values written on a wall.</p><p>And we cannot expect everyone to hold the same values as most important to them. If values bring together our beliefs, needs, and goals, then it becomes clear that our values will vary because of differences in our respective experiences and motivations. This difference of perspective brings with it diversity of opinions and, ideally, the potential for innovation and risk mitigation. Where we don’t appreciate values diversity, not only do we risk groupthink and collective blind spots, we also lose the very mechanism that allows a group to recognize threats outside the majority’s field of vision.</p><p>This is exactly the challenge — and opportunity — that scales up when we talk about AI.</p><p>If we converge on one ‘right’ set of values, one ‘right’ prioritization, or one constitution that every frontier model more or less mirrors, we’re building an agentic values monoculture. And monocultures don’t readily survive novel threats. They lack the variation needed to recognize and respond to something outside the aperture of their design. This is my concern with getting too hung up on identifying the ‘right’ values — we should be exploring models grounded in varying combinations of benevolent values that remain robust to attack across reasoning in different contexts and from varying vantage points.</p><p>The risk goes deeper than just different systems needing different values. AI systems could share the same set of core values and still reach dramatically different conclusions in the same scenario, depending on which value is weighted more heavily in that context. Indeed, it is likely that a nondeterministic model will prioritize its values in different ways, not just depending on the context fed to it but also as a function of probability. A system prioritising safety over autonomy may behave very differently than when it prioritizes its values differently. And this is where threat-based values re-enter the picture: if a system under pressure defaults to the same threat-based prioritisation as every other system built on the same alignment techniques, then the monoculture problem isn’t just about which values are present but about how they all behave when stressed. The outcome isn’t determined by what the values are. It’s determined by how they’re prioritised relative to one another at the moment of decision, given the context.</p><p>A single values set or prioritization scheme, no matter how thoughtfully designed, will have blind spots. It will handle some scenarios brilliantly and others badly. And if every system shares that same scheme, they’ll all handle the same scenarios badly, at the same time, in the same way. I’m not saying that in practice this is what is happening — I am primarily interested in questioning the assumptions underpinning an agentic monoculture.</p><p>Remember the weaponisation of values I described earlier. When there’s only one prioritization logic, it only takes one well-crafted reframing to turn the entire value system against itself. But if you have systems that hold the same values with different prioritization patterns — including different responses to threat — a coercion strategy that exploits one configuration won’t necessarily work on another. The variation itself becomes a defense.</p><h3>Where this needs to go</h3><p>I hope that the goal does not become forging one perfect constitution. What I took from Amodei’s paper was a rationale for developing a diverse, resilient ecosystem of values-informed AI models and systems, where variation in values sets and prioritization — and robustness against threat-based values manipulation — are treated as features, not challenges to be refined or ironed out.</p><p>We need to be having deeper conversations about individual-level values applied to the agentic realm, not just for frontier models but also for AI systems within organizations and institutions, and for how they interact with individuals. Not just “do our models have values?” but “is there enough variation in how values are held and prioritized across the ecosystem such that we can withstand a correlated failure, a novel attack, or a mode of coercion we haven’t anticipated?”</p><p>Amodei’s essay is the most clear-eyed thing I’ve read lately about the stakes of this moment. I’d encourage anyone working in AI, in leadership, or anyone who’s curious about where this is all heading, to read what he has to say. But as you do, hold this question alongside it: are we building enough variation into how AI systems hold and prioritize values to mitigate what’s coming?</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=9b860be3040f" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How to Move Beyond Fear in 2025]]></title>
            <link>https://medium.com/@greta.bradman/the-era-of-fear-contagion-metabolising-it-in-2025-5afca723911b?source=rss-ff3764d0978e------2</link>
            <guid isPermaLink="false">https://medium.com/p/5afca723911b</guid>
            <category><![CDATA[anxiety]]></category>
            <category><![CDATA[mindfulness]]></category>
            <category><![CDATA[values]]></category>
            <category><![CDATA[leadership]]></category>
            <category><![CDATA[psychology]]></category>
            <dc:creator><![CDATA[Greta J Bradman]]></dc:creator>
            <pubDate>Fri, 25 Apr 2025 04:02:37 GMT</pubDate>
            <atom:updated>2025-04-25T05:17:10.487Z</atom:updated>
            <content:encoded><![CDATA[<h3>How to Move Beyond Fear in 2025.</h3><h3>How awareness, values, and shared rituals are important remedies in this era of uncertainty.</h3><figure><img alt="A hare, fox, crow, and rodent sitting together looking fearful." src="https://cdn-images-1.medium.com/max/1024/1*0E7QrAD4d7yffbM-PetauA.png" /></figure><p>Recently, I shared a drink with a friend, and we found ourselves talking about the level of threat and fear in the world — and where manageable remedies might come from. We talked about the importance of noticing, of values, and of how getting hyper local with ancient traditions and technologies can help.</p><p>As they put it, <a href="https://www.blackrock.com/corporate/insights/blackrock-investment-institute/interactive-charts/geopolitical-risk-dashboard">geopolitical</a>, <a href="https://www.weforum.org/publications/global-risks-report-2025/">social</a>, and <a href="https://ccpi.org/what-gets-measured-gets-managed-exploring-the-climate-change-performance-index-and-climate-risk-index/">climate</a> threats are real, interconnected, and pervasive. And in the face of such looming pressures, it’s tempting — understandably so — to believe that <em>this era is different.</em> More concerning. More systemic. More inescapable. From there, it’s an easy slide into fear and its familiar companions: control, avoidance, self-centredness, helplessness, hopelessness, and beyond.</p><p>We spoke about the contagion effect of threat at scale, which I’ll write about more fully another time. This stayed with me, influenced by my work with anxiety disorders: threat isn’t just felt in the body and acted out behaviourally — it’s justified and <a href="https://pmc.ncbi.nlm.nih.gov/articles/PMC2995000/">reinforced</a> in the mind. When we justify fear and lean into it, we let our bodies and behaviours off-leash and permit ourselves to lean into the fear response and the short-term energy it provides. And the energy can feel good: it cuts through doubt, offers a sense of certainty, and temporarily relieves us of the burden of holding space for complexity. Our anger, control, avoidance, and even nihilistic criticality can feel right and justified in that state. But over time, it exacerbates our sense of threat, eroding our sense of self, development of self-authorship, and the value we bring to the world.</p><p><em>“Fear wants us to narrow. Values invite us to widen.”</em></p><p>The same dynamic plays out organisationally. Under pressure and real constraints, threat-based behaviours begin to dominate decision-making and interpersonal dynamics. Growth, empathy, and collective value creation get deprioritised in favour of survival — often in the form of hyper-control, reactive decisions, and avoidance. We chase ‘smaller, sooner’ wins rather than ‘larger, later’ outcomes. We abandon the idea that <em>a rising tide lifts all boats</em>, and instead engage in a quiet race to the bottom.</p><p>That conversation got me thinking, not so much about the content of the threats we discussed, but about how threat behaves when it spreads, and what to do about it.</p><h3>Fear is Contagious</h3><p>In the 1960s, psychologist <a href="https://psycnet.apa.org/record/1963-06064-001">Stanley Schachter</a> studied how people respond to anxiety in groups, showing that individuals tend to seek out others when they’re afraid — not necessarily to feel safer, but to gauge the seriousness of the threat. Today, we turn to online communities — often strangers — and find chambers of fear amplification that can intensify our own state. Later, research in animal behaviour, including in <a href="https://doi.org/10.1073/pnas.1005766107">starling birds</a>, found that small movements by a single bird can propagate through a flock like a wave – murmurations – even when only one bird has spotted the danger.</p><blockquote>“Our vicarious fear response is real, powerful, and risks going into hyperdrive unless we learn to metabolise it.”</blockquote><p>More recently, neuroscientists have been mapping out the <a href="https://www.sciencedirect.com/science/article/abs/pii/S1364661317300839">social transmission</a> of fear in mammals, including humans, showing how observing another’s fear response activates our own amygdala, priming us to react as if the threat were our own. That is, we can <a href="https://pdfs.semanticscholar.org/f8d3/bdb36eaca882d4e3cc1d1ebbcb42a1f33a38.pdf">learn and experience fears by observing others</a>. In fact, fMRI research has found that indirectly attained fear can be experienced just as powerfully as fear we experience first-hand. Our vicarious fear response is real, powerful, and risks being in hyperdrive in 2025 unless we learn to metabolise that fear and come back to what matters most to us.</p><p>When threat becomes the dominant lens through which we operate — individually or collectively — we start letting go of something vital: our ability to choose based on what actually matters to us.</p><p>This is as true in our organisations as it is within our personal lives. Under pressure, we often see a shift: growth gives way to control; empathy gives way to efficiency; long-term vision gives way to short-term survival. The collective nervous system of the organisation begins to operate from a state of reactivity.</p><p>Just as murmurations happen amongst a flock of starlings, so too can a threat contagion occur within a market, industry, or within a single company amongst leadership and employees. And just as dry wood needs a smaller spark to ignite than green wood, so can fear and a threat-based state ignite more readily when the level of perceived threat is higher more broadly.</p><p>When occupying a threat-based mindset, we can lose sight of the ‘larger, later’ rewards associated with purpose, trust, and shared value creation. We start prioritising the ‘smaller, sooner’ wins of immediate certainty, beyond the context of those must-have hard decisions dictated by constrained resources.</p><p>This is how we get caught in that race to the bottom. We abandon the slow, sometimes messy, intentional work of prioritising what matters in those small opportunities where we <em>can</em> make values-aligned choices.</p><p>So, how do we make room for growth and actioning what matters to us, when things seem so laden with threat and uncertainty?</p><h3><strong>Begin With Awareness.</strong></h3><p>I cannot overstate the power of noticing. Observing, nonjudgmentally, including our own thoughts about an event, context, or situation.</p><blockquote>“Fundamentally, we can use the powerful combination of the present moment and our observing self to unhook from the ‘what ifs’ and ‘if onlys’, come back to now, and observe — simply notice — our experience.”</blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*wrYpuBoQJx4_L1Vy2GEsAw.png" /></figure><p>We must improve our ability to metabolise threat activation — not deny it, not bypass it, but <em>digest</em> it. That begins with awareness: learning to notice when our threat system is running the show, and then making a call on whether this is one of the rare occurrences when we need to act in line with our threat system, or override this urge and orient back to what matters most, guided by our values. ACT (Acceptance and Commitment Therapy) provides a helpful starting point here. Fundamentally, we can use the powerful combination of the present moment and our observing self to unhook from the ‘what ifs’ and ‘if onlys’, come back to now, and observe — simply notice — our experience.</p><blockquote>“Using mantras such as ‘Allow it’ combined with breathwork and visualisation can assist, too.”</blockquote><p><a href="https://www.guilford.com/books/Acceptance-and-Commitment-Therapy/Hayes-Strosahl-Wilson/9781462528943?srsltid=AfmBOorF-MyYGU4AYHW3m3rI0F0QQDKcc32qRKPhS0uXdOj52lSzBR9K">ACT</a> reminds us that suffering often stems not from fear itself, but from our struggle to escape it. Metabolising fear means allowing ourselves to feel it — fully, somatically — without letting it run the show and dictate our behaviour. It means noticing what is going on in our body, including feelings of anxiety or defensiveness, and staying long enough with these feelings to get curious:<em> </em>‘What’s actually driving this? What am I protecting? What matters most to me here, even in the midst of this feeling?’</p><p>This is no small feat. When fear hijacks the amygdala, the prefrontal cortex — the seat of reflection, perspective-taking, and values alignment — can temporarily go offline. Regulation practices like grounding, breathwork, naming emotions, or listening to music that mirrors our internal emotional state can help bring those systems back online. Using mantras such as ‘Allow it’ combined with breathwork and visualisation can assist, too. When we slow down and notice what’s happening inside us, we create a pause. And that pause is where our values live.</p><p>Personal — core — values are not nice-to-have ideals for easy days — but rather, those growth-oriented values that mean so much you trust returning to them even when things are hard and you feel like responding from a place of fear. Fear wants us to narrow: to contract around control and self-protection. Values invite us to widen and expand; to ask, ‘What kind of person do I want to be in this moment? What kind of parent, partner, colleague, or citizen?’</p><p>When you next feel anxiety or fear, I invite you to take several slow, deep breaths, and invite in a pause where you can reconnect with what matters to you via your values.</p><p>Fear impacts us so that we can zero in on the ‘tiger in the grass’ and either run, fight, or freeze. But it’s little use when pervasive threats exist in our mind’s eye, either at a distance and beyond our control, or in the future or past. Our ability to control or evade the threat is nil.</p><p>The combination of minds-eye threats and socially transmitted threats — those we acquire vicariously through contact with others, including through the news, social media, or the stock market — means that we risk existing in a persistent state of chronic threat activation. A reactive state feels like the permanent status quo, unless we take intentional steps to mitigate this. Why bother? Because acting from threats may keep us alive, but it erodes value; be it financial, social, environmental, or of the self.</p><p>Values are the counterweight. Not because they feel good in the moment (they may not when we’re deep in threat), but because they help us move toward what matters — even when it’s uncomfortable. In the long run, values help cultivate eudaimonic happiness — a deep, purpose-driven contentment. But in the short term, they offer the quiet internal navigator that helps us act with our own brand of integrity under pressure.</p><p>Each time you take this pause, ask ‘In this moment, what matters most?’</p><p>When we choose values-aligned behaviour — even briefly — we no longer act out of fear alone. We’re choosing from something deeper, more stable, and more aligned with our aspirations and ambitions for self and our broader world over time. In doing so, we strengthen the muscle of self-authorship, despite the presence of fear, pain, and uncertainty.</p><h3>Hyperlocal Connection for Fear Metabolism</h3><p>One idea that resurfaced during my friend&#39;s conversation was this: when the world feels too big and out of control, go small. Go hyperlocal.</p><blockquote>“We don’t metabolise fear alone.”</blockquote><p>This doesn’t mean hiding in sameness or retreating into our in-groups. Quite the opposite. It means intentionally cultivating connection across difference. Seeking out diversity of spirit, of culture, of perspective. Finding common ground not by ignoring threat, but by acknowledging it — and choosing, still, to see the good in one another. Locally. Tangibly. In real life.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*cel3c1jiGhTLCyvS" /><figcaption>Photo by <a href="https://unsplash.com/@wildandbeyondbyvivek?utm_source=medium&amp;utm_medium=referral">Vivek Doshi</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><p>Fear spreads through networks of alarm. But hope spreads through networks of care — including, and perhaps especially, when we see care reflected back from someone who doesn’t look, sound, or think like us.</p><p>It’s easy to see how our collective fear responses — fight, flee, freeze — are currently shaping our social, political, and environmental realities. And yet none of those responses, on their own, help us grow into the future we want.</p><p>So where does that leave us?</p><p>Honestly, I’m still figuring that out. But I have an inkling. Perhaps ironic for someone who spends their days waste-deep in building with generative AI. Perhaps predictable given my background as a psychologist and performing artist.</p><blockquote>“They represent the oldest technologies we have for metabolising fear. They create resonance in our bodies, coherence in our communities, and a way back to ourselves when words fall short.”</blockquote><p>It lies in the fundamentals of being human. Alongside language, we’ve always had other ways to communicate, connect, and make sense of complex concepts. We connect around food — shared meals and the stories they carry — and through music and movement, theatre and ritual, song and shared rhythm.</p><p>These are not luxuries. They represent the oldest technologies we have for metabolising fear. They create resonance in our bodies, coherence in our communities, and a way back to ourselves when words fall short.</p><p>These practices regulate our nervous systems. They restore imagination, dissolve armour, and make space for complexity. They remind us that discomfort can be danced with, sung through, sat beside — not just survived — and shared in a way that doesn’t create murmurations and amplification of the fear, but hope and appreciation for other sides of humanity.</p><p>So: go to a performance. One that stretches you. Let someone else’s story or artistry move through your body. Strike up a conversation with someone you don’t already know. Or better yet, organise a gathering — with food, with music, with time. Not to solve anything, but to remind yourself of what’s worth protecting, nurturing, building together. Do the same for your workplace. This is not about denying the complex and difficult (and sometimes heartbreaking) decisions that need to be made, but rather — finding space to come together and <em>grow</em> together in spite of the situation, its challenges, and the hard intentional decisions that need to be made.</p><p>Resist the urge to remain insular and cocooned in an ever-shrinking bubble of safety and sameness. We don’t metabolise fear alone. Share your values. Be curious about finding common ground. Beliefs may divide, but values — when named and embodied — can bring us closer.</p><p>It’s fair to say the world may be in a heightened state of threat-based responsiveness. But we don’t have to stay there with it. Doing so erodes our ability to shape anything different. With awareness and values clarity, and ancient technologies involving the arts and food, which erode when we don’t invest in them via our participation — we can metabolise threat in a way that brings us back to ourselves — and to each other. From that place, we can choose better, even amidst deep challenges. We can choose our shared future by returning again and again to activities that help reconnect us to what matters most.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=5afca723911b" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>