“The human species thinks in metaphors and learns through stories” — Mary Catherine Bateson

As someone who’s spent most of my time in the natural sciences, metaphors and analogies did not come naturally to me. Yet, our vibrant digital ecosystems are teeming with them: ‘Data is the new oil’, ‘Technology is a just a tool’ or “Internet is the information superhighway”. What did not occur to me was that these metaphors were inadvertently shaping not only the way I saw and understood technology — but also how its designed and used.

As our technological world grows increasingly more complex and sophisticated, we are beginning to see not just a digital disruption across our many sectors — but the consequences of how we’ve chosen to think about technology.

Take for example, “Data is the new oil” — while there have been many articles describing the ways in which data differs from oil, such as its unlimited supply, how it can be reused and add value — what’s missing is the deeper meta-messaging around its use. That data can be ‘extracted’ as a commodity (similar to oil) carries capitalist connotations, which taken too far, could result in a violation of human rights depending on the data.

Language around ‘data mining’, ‘data pipelines’ and ‘data engineering’ invoke imagery related to the Industrial era — one of power-driven machinery and various refining, transportation or warehousing processes. Indeed, even the methodologies, tools and techniques developed during this time were focused within that context. One example being the use of statistical tools to improve production or manufacturing processes (my favourite modern version is the coffee valve defect rate study!).

But embedded in this setting come ideas from other entangled contexts relating to the social and economic systems of the time. And where previously the focus was on defective machine parts and improving production efficiency, these metaphors and analogies have been creeping into the way we work with human data with the stated aim of improving ‘human productivity’.

We need to stop and ask the question — “What were the original intention and purpose behind the development of these tools and techniques?” and also “Are those same tools and techniques appropriate or applicable in the same way for human data?”

There’s Data… and then there’s Data

“We have to be careful of the metaphors we use to make meaning, because metaphors are the language of spirit and that’s how we operate in our fields…”

— Tyson Yunkaporta from “sand talk”

For all the classification methods and taxonomies we’ve developed for data, few have really considered how this information collected has been decontextualised in the process of measurement or recording — and more importantly, what this loss means in terms of the limitations or validity of the tools we use.

If we were to use the metaphor to our advantage — and change it to “data is the new apple”, how would we think and work differently with data? For a start, it’s worth recognising that the apple comes from a tree, which belongs to a forest and is in relationship with the soil and creatures in that forest.

Perhaps we could reflect that some data cannot be taken out of its many contexts without losing some of its quality.

Take for example, Artificial Intelligence (AI) algorithms that are trained on human responses in a busy city may differ vastly from how it would perform in a rural village in India. Or that certain types of data may have a ‘shelf-life’ — data from the 1950s on women in the workforce may not really be appropriate as input for predicting work performance for women today.

Obvious? Perhaps.

But while these examples seem fairly straightforward, what is not clear is where the ‘delineation’ that separates information that can be used from that which shouldn’t lies.

And what makes this even more complex is that the many entangled contexts, for example around societal attitudes, racial justice and economic equality, affect some types of data more than others.

So it’s difficult to really know which contexts affect the viability of our data for a specific purpose. Nevertheless, it’s important for us to not only be aware that they might — but also have an awareness of the consequences of these limitations.

Yet, organisations still use human data to predict human behaviour, such as the propensity for an individual to leave their organisation, without much attention to not only the ethical implications for their employees — but also the analytical rigour in assessing the viability of the models used. Meanwhile, others are discovering that even large datasets and the ‘best predictive AI-models’ do little to predict life outcomes for social programs and in the criminal justice system.

Technology is a Forest and the Internet a River

I’ve often heard arguments about the ‘neutrality’ of technology accompanied by the metaphor that “technology is just a tool” — which in my mind always brought up the image of a person holding a hammer looking at most things as a nail!

What I’ve been oblivious to, until now, was the relationship between the person, the hammer and the environment — and how the attributes of that hammer change the perceptions within the person and their relationship to their environment… And we haven’t even got to thinking about how this applies to AI yet!

Questions around how these ‘tools’, such as AI, shape our perceptions and change our inter-relationships within society, and our systems such as health, education, economic, legal etc., depend not only on the tool itself — but also our human ability (or inability) to foresee the totality of the consequences in using them.

I prefer the metaphor, “Technology is a Forest” because it conjures up a deeper sense of awe within me — a sense that reminds me to be humble, lest I disrupt the larger ecosystem by lighting a small fire only to create an uncontrollable bushfire!

Compare that to the fire that Apple came under when they launched their credit card in August 2019, without really knowing how their AI credit rating algorithm managed to evade third party vetting to discriminate against women. It’s a tough call, but digital fires may travel faster and further than their real counterparts…

Then there’s the “Internet, the Information Superhighway” - but these days it’s more like a river that’s turned into a swamp of fake news and misinformation. Yet, most of the public debate on the regulation of social media and messaging apps are focused on regulating the content on these platforms — so it’s like making sure the ‘traffic’ on these ‘information superhighways’ are ‘roadworthy’…

But I prefer the ‘river’ metaphor — we need to stop and ask, “What is it about the conditions in and around this river that’s turning it into a swamp, where the fish get sick?” and “What are the conditions required for digital platforms to become healthier places to interact virtually?

It is in choosing our metaphors wisely, that we may be able to better intuit and understand the technologies around us, how we should be using it more responsibly and how it may be designed and regulated. So the next time you think of a tech-metaphor, it might be worth pausing to reflect on what that says about the way you think about technology and society!

“The way in which a concept is delivered speaks not only to the concept itself but also offers a description, between the lines, of what is unquestioned.”

— Nora Bateson from “Small Arcs of Larger Circles”

— — —

Phoensight is an international consultancy dedicated to supporting the interrelationships between people and technology, and is accredited by the International Bateson Institute to host Warm Data Labs.

Why Metaphors are Important in Technology was originally published in Phoensight on Medium, where people are continuing the conversation by highlighting and responding to this story.

Open Source Software (OSS) first emerged as a social movement (along with the free software movement) in response to the economic and legal constraints in the early 1980s. Its evolutionary leap had deftly subverted the market’s ‘invisible hand’ — and was spurred on by the reticence of those unwilling to be beholden to the restrictions and intellectual property rights of proprietary software.

The philosophical underpinnings of this new breed of software were about accessibility, freedom of use and a spirit of collaboration. And the idea that computer software could succeed in a competitive market if it was freely available to use, share and modify was previously unheard of.

What’s fascinating about the rise of OSS is the novel way that software or ‘information’ could evolve, flow and be controlled in our technological landscape within the existing economic and legal structures in society.

For the most part these structures pre-dated the digital era which brought on (essentially) costless digital copying, near real-time information transfer and distributed, dense connectivity across geographic boundaries. OSS was not your regular ‘market product’ with characteristics that were familiar to society at the time. Yet these new digital attributes augmented with human capability demonstrated a new way around old structures.

OSS disrupted existing monopolies by de-centralising power and providing the transparency that developers required for better technological solutions. Its licensing helped ensure its integrity through verified source validation — and this also served to strengthen trust in the software within the community. As we’ll see later, ‘objective’ verification of software plays an important role in the growth and scalability of OSS.

With a large, dedicated open source community that nurtured its development and the protection of a ‘Copy-Left’ licence, OSS had the conditions it required to thrive. Its strength lay within its developer base and its mutability, which allowed it to adapt to different contexts as required. This ability — that of being able to build on or modify existing software for bespoke code — is another important feature we’ll shortly revisit.

Objectivity — A requirement for Scalability

With early OSS, what was shared was software or information that could be called ‘objective’ — and by this I mean that programmers and other developers who read the code could mostly agree on the ‘coherence’ between the software and what it returned when it was executed (assuming it didn’t crash or hang!). In other words there was a certain ‘reliability’ about this software, in so far as the interpretation of the code with its corresponding output.

The information communicated through the code was stable (compiler issues aside!), and did not incur much loss of meaning or contain any additional information through the ‘transmission’ between developers. Further, the openness of OSS allowed programmers a deeper understanding of its functionality and how it could be modified.

What this meant was that it could be successfully scaled — because it was trusted to ‘behave’ in a way that was somewhat predictable (with mostly known consequences) and OSS developers could modify the code to suit their specific needs.

The ‘objective’ nature of this kind of software lends itself well to mechanistic analogies — like an engine with many interconnected parts. With such software, you could change the parts (or fix any bugs!), put in your own code or remove parts to create a new ‘engine’. You could modify existing OSS for your own purpose!

This ability to adapt to bespoke needs gave OSS a new kind of resilience in its functionality — it was also more responsive to external requirements mainly due to the flexibility, versatility and ingenuity of OSS developers. Let’s call this kind of adaptability a ‘first order modification’ — where parts or code (information) can be replaced, added or omitted, but can’t autonomously change themselves or mutate.

So far, all is well — but what if the software we were creating was more volatile? What if it behaved more like an organism than a machine — with an ability to respond to and learn from its ecosystem? Where the parts themselves could autonomously change and mutate? Let’s refer to this type of adaptability as a ‘second order modification’.

What if its functionality depended on the communication of information that was highly variable (i.e. where the variability between the information sent and the information received was large)? What if there were intrinsic losses in the information being transferred? What if the software performance depended on its relationship to the local conditions of where it was being transferred?

How would these affect the development and scalability of software?

We are now arriving at a new threshold — one where ‘objective’ information is replaced by ‘subjective’ information, and is therefore context dependent. We are entering the forest of living systems and multiple contexts...

Information Crossing Thresholds

The philosophy behind ‘Open Knowledge’, similar to that of free software, is that any information that contributes to the the betterment of society should be freely available — and free from any barriers or constraints that hold back human progress. Implicit in this understanding is the assumption that this knowledge or information is ‘stable’ and ‘lossless’ during the distribution or transfer.

And it is here that I want to pause, and explore a few aspects of the nature of ‘knowledge’ — aspects that examine the ‘factual’, objective side of knowledge, but also its ‘experiential’ or subjective side.

In an Industrial world, where objective ‘facts’ move freely and are mostly unchanged across many contexts, scalability has offered many benefits to society — the mass production of transportation etc. Subjective information or knowledge, however, is highly context dependent — information about experiences of how a community survives in the Amazon rainforest may be very different to that for the Arctic, and may depend on the inherent characteristics and culture of that community!

Distributing such (subjective or localised) knowledge without its rich and many contexts does not only render the information being transferred as ‘leaky’ — it may actually be harmful if applied out of context.

Anecdotally, many indigenous cultures recognise the importance of the multiple contexts related to information, and are careful that the process of distributing knowledge does not violate its integrity.

The trick is knowing whether we have a ‘viable’ information set (or code) which also contains information about the many contexts around it so that it’s useful and reliable. Obvious? Maybe not!

With the digitisation of information and the increasing reliance on subjective or personal data — many ‘Artificial Intelligence (AI)’ algorithms that are being developed and distributed are trained on data that’s ‘de-contextualised’. What this means is that some of the information has already been ‘lost’ in the process of collection — with algorithms being trained without the necessary contexts for it to be viable for certain uses.

And this is not easy to ascertain! How does one even determine which contexts are relevant for use if one does not know how the algorithms will be used after distribution?

Moreover, some of this data is subjective information that’s formed through someone’s unique perspective or behavioural response, and received (or interpreted) through another different lens. In fact, it’s much more complicated than that — what if that information was also shaped by localised economic and legal structures? Would a ‘first order modification fix’ of these algorithms be sufficient to make it suit localised needs?

How do these algorithms, that are shared or distributed, ‘unlearn’ data from those earlier contexts that are no longer relevant? How does one untangle the contexts of ‘culture’ or ‘time’ from data? Human memories (unlike digital information) shift through many different perspectives with time. The complexities that surround distributed algorithms today are wide and varied — but what’s important is that they maintain their integrity.

In the same way that ‘copy-left’ was able to control the ‘flow’ and ‘integrity’ of code within existing economic and legal structures, perhaps what could also emerge are other variants of licensing or supporting mechanisms for ensuring the responsible use of algorithms despite the current lag in regulation.

Perhaps if there’s one thing I’ve learned about the history of Open Source and Free Software it’s that human ingenuity could once again be used to protect the integrity of algorithms, so that we can be more rigorous in how we use, create and deploy new generations of software — including AI — in society.

“It is very important to examine how you reach a conclusion, and what premises it depends on, to see where it might be valid.” — Richard M. Stallman

— — —

Phoensight is an international consultancy dedicated to supporting the interrelationships between people and technology, and is accredited by the International Bateson Institute to host Warm Data Labs.

Open Source, Scalability and the Forest of Contexts was originally published in Phoensight on Medium, where people are continuing the conversation by highlighting and responding to this story.

The emergence of new generations of Artificial Intelligence (AI) algorithms, (such as GPT 3), is creating an urgency in addressing the socio-technological issues around AI. Along with all the hype comes unprecedented challenges on how organisations can manage the repercussions and risks of deploying AI technologies — and who is accountable when things go wrong…

In fact, a recent survey of IT and Business Executives across nine countries (Australia, Canada, China, France, Germany, Japan, the Netherlands, the United Kingdom and the United States) indicates that 56 per cent of organisations are slowing down their adoption of AI technologies due to concerns around emerging risks and negative public perception.

While some are ‘finely tuning’ AI algorithms for ‘near-perfect performance’ in testing environments, the recent example of the gender bias with Apple’s credit card rating likens their deployment in society to a bull in a china shop! Not to be taken lightly when so much is at stake, AI presents broad and complex challenges that are unprecedented compared with previous technologies.

Organisations are caught between a rock and a hard place — do they play it safe and forego the touted “economic growth and competitiveness”? Or do they take a calculated risk that could potentially damage their reputation and trustworthiness — not to mention any unforeseen collateral damage?

The recent ‘State of AI 2020’ report notes an alarming and rapidly approaching growth in the computational, economic and environmental costs for ‘incrementally smaller improvements in model performance’. At a time when organisations are committing to transitioning to a ‘net zero carbon emissions’ future, questions around governance — when and how to use AI — will make their way into the boardroom.

Meanwhile, various AI Ethics principles and guidelines are being developed with the intention of restraining the potential harms wrought onto society — but the research indicates that these have little impact on human decision-making. Calls for auditing frameworks, more algorithmic transparency and explainability, while incredibly important, don’t address the hidden social and ecological harms — most of these aim to flag or manage the manifestations of AI harms that can be detected rather than those that are imperceptible.

In amongst all these complexities, board members, senior leaders and chief data officers are expected to foresee the potential risks and socio-technological vulnerabilities for better decision-making. With so much uncertainty, it’s no wonder that many are pressing governments for AI regulation so that some line may be drawn in the sand.

The task is daunting, but there are few things we can do at the outset. First, we need to take a step back — away from fixing the plumbing in the AI pipeline — to survey the changing terrains and navigate our way forward by asking three questions:

1. How are our Perceptions changing?

“Perception is crucial to understanding. How you see, and what you see, determine how you will be… ”

— John O’Donohue

The fundamental quest of big data and AI is the discovery of new insights into existing organisational challenges and the many issues we face today. Yet, how these issues and challenges are framed largely depends on how we perceive them — and this ‘framing’ extends to the data we collect and use, which in turn shapes the AI algorithms that are employed.

Our perception of the problems and the information we use to better understand them are crucial in setting up our strategy (AI or otherwise) forward. While AI algorithms can be incredibly useful in some instances, they suffer from an inability to adapt to multiple contexts in the way humans can — Why? it’s simply because the data they use is de-contextualised.

Now, this can be both an advantage and a disadvantage — in cases where the information is ‘objective’ (or has little variation with changing contexts e.g. normal human body temperature) AI and other computational processes are less likely to ‘go off the rails’, but these powerful tools can prove to be a huge liability when using ‘subjective’ information where context matters!

And we’ve seen examples of this in society — discriminatory algorithms for job screening, falsely accused criminal charges, and increased inequality for credit ratings are just a few amongst many. But public debate around these issues still mainly focuses on data biases, and whether or not it’s possible for these to be mitigated by ‘curating’ the data or redesigning AI algorithms.

What’s often lacking is a deeper understanding of the significance of the many missing contexts surrounding the data — and a realisation that while some data can appear to be ‘objective’, systemic biases within society taint this information, and has the potential to perpetuate or exacerbate the problem.

So we need to start by asking new questions.

What sort of data are we collecting? Whose perception does this information represent? What are the contexts that are missing? How does this change our perceptions? How do we interpret any data-driven insights across multiple contexts?

2. Where is the Knowledge?

“Warm Data is not meant to replace or in any way diminish other data, but rather it is meant to keep data of certain sorts “warm” — with a nest of relations intact.”

— Nora Bateson

In a competitive market where the knowledge economy is fuelled by a race for information that allows organisations to become more productive and efficient, the use of AI algorithms is typically seen as a sign of technological progressivity. As sophisticated AI algorithms, such as GPT-3, become more costly to train and run, the market dynamics between small businesses and larger companies are also expected to change.

While it may appear that the funding gap disadvantages the smaller players, what’s been overlooked is the type of knowledge and information that’s available. Organisations have long known that ‘tacit knowledge’, though difficult to quantify, can play an important role in understanding business and client needs. The challenge has been in finding ways to better harness and utilise these ‘human assets’.

Once again, a deeper comprehension of the role of ‘inter-relationality’ and the many contexts involved could unearth new knowledge that has the potential to rival scalable AI technologies by applying localised and nuanced insights for a competitive advantage. Digital transformations that are careful in how they value these human assets, particularly as to the contextualising of information, have the potential to ‘leap-frog’ more blunt approaches with better outcomes.

Some questions to consider might be:

Where is the critical knowledge we need? What insights and learnings can be gained by examining the various inter-relationships within, across and connected to our organisation? How do these relate to more traditional knowledge sources?

3. How do we Respond?

“The ethics of the system are the intrinsic aim of the system, not an externally imposed restraint on commercial or other outcomes.”

— Worimi man Deen Sanders & Kamillaroi man Rick Shaw

Strong organisational culture and values are critical for anchoring client trust during uncertain times. Along with the recent global issues such as climate change, increased inequality and political polarisation, we are witnessing a shift in public attitude and awareness when it comes to corporate responsibility, accountability and sustainability.

Investors are being forced to diversify their investments away from systemic risks, by considering the environmental and social impacts as part of their fiduciary duty. In the same vein, companies are recognising that employee satisfaction is critical to good performance. In amongst all these shifts come calls for increased AI accountability and responsibility.

As we build for a new future, the social costs of digitisation, as they affect human rights, privacy concerns and the vulnerable, bring to light complex challenges on how to navigate the ethical dilemmas that come with designing AI algorithms. Indeed, tech employees are increasingly taking on the additional burden and moral responsibility in tending to the ethical risks of technology with little support.

At a time when organisational resilience has never been more important, it may be prudent to re-assess whether digital technologies are being used for greater efficiencies at the cost of employee and service vulnerability.

In his work on better understanding systemic resilience, French geophysicist, Xavier Le Pichon, puts forward the idea that tending to the vulnerabilities within a system are important for facilitating its evolution, and for increasing its ability to adapt during times of uncertainty.

Could certain ‘perceived’ inefficiencies actually be advantageous during turbulent times? Could understanding the many contexts and nuances of the complexities we are working within allow us to find new ways of using the technology — without compromising our values or moral responsibilities? How can we ensure a robust corporate governance that supports a healthier workplace culture and ecosystem?

“Human culture is a decentralized evolutionary system… Any predictive model that fails to incorporate this distributed ongoing daily billion-headed microevolution is doomed to collapse…” — Kevin Kelly

This article is based on my position statement at the 5th IEEE UV2020 panel on “New legal, social, and ethical challenges posed by applications of AI” and also my presentation, “In the Shallow with AI” at the Big Data Conference Europe 2020.

We will be hosting a series of workshops on how organisations can navigate the complexities presented by AI in early 2021.
Register here for more information.

Phoensight is an international consultancy dedicated to supporting the interrelationships between people, public policy and technology, and is accredited by the International Bateson Institute to conduct Warm Data Labs.

Three Questions you need to ask before you use AI in your organisation was originally published in Phoensight on Medium, where people are continuing the conversation by highlighting and responding to this story.

“A world perfectly fair in some dimensions would be horribly unfair in others.” - Kevin Kelly

“Fairness” in Artificial Intelligence (AI) applications — both as a concept and a practice — is the focus of many organisations as they deploy new technologies for greater effectiveness and efficiencies. That machines are faster at processing large amounts of information and the notion that they are ‘more objective’ than humans, appear to make them an obvious choice for progressivity and seemingly impartial actors in ‘fairer’ decision-making.

Yet, algorithmic based decisions have not come without their share of controversies — Australia’s recent ‘robo-debt’ government intervention which wrongly pursued thousands of welfare recipients; the UK’s ‘A-Levels fiasco’ of downgrading graduating grades based on historical data, its controversial visa application streaming tool; and concerns about Clearview AI’s facial recognition software for policing are raising new questions on the role of these technologies in society.

Risk assessments are part of the fabric of modern society, but what we are dealing with here is not just ‘scaling up’ human capacity for decision-making without the unwanted human biases and errors — we are also extolling the ‘virtues of objectivity’ under the guise of ‘fairness’ (which is inherently subjective!) and failing to recognise the many inter-relationships that are being unraveled through the use of these algorithms in our daily lives.

And it is these inter-relationships that are holding together the systems we find ourselves in.

Issues around ‘trust’ when it comes to AI are multi-faceted — explaining how these technologies arrive at a particular outcome in decision making and its reliability is just part of the story! Understanding how our societal systems will change as a result of AI goes to the ‘trust’ that is within the relationships between people and their changing world.

What we have really created is a way to disrupt these old structures and the perceptions on which they were founded. AI and other decision-making algorithms are forcing us to revisit the moral underpinnings of how we think about ‘fairness’ and its role in our society and where the trust really lies...

Facing up to AI

“Against the infinity of the cosmos and the silent depths of nature, the human face shines out as the icon of intimacy” — John O’Dohohue

The use of facial recognition in AI has received much attention in the media, particularly when it comes to human rights and privacy. In a recent article, the New York Times covered some of the many risks in using facial recognition technology including: its reliability and limitations; how it’s implemented and used; and the legal and moral challenges faced by society in navigating this ethical minefield.

Calls for more transparency in AI applications, though critically important in understanding hidden biases and uncovering the underlying values in the algorithmic design, only scrape at the surface of deeper societal issues around justice and fairness.

To view ‘transparency’, in this context, as a window into the decision making process from data to output, is to miss how interfacing AI with society is altering our current systems.

No amount of ‘band-aid solutions’ to either the algorithms or the underlying data will be sufficient in addressing what are inherently system-wide problems.

AI transparency and the toolkits and guidelines for building trust in these technologies do not go far enough in providing insights into how our systems are being affected across many different contexts such as social, economic, financial, political and educational amongst others.

Take for example, using AI to analyse a job applicant’s facial movements to determine their suitability for employment in a particular industry or the claim that AI is able to predict a job applicants propensity for job-hopping — both which evoke sentiments similar to early eugenics!

While the algorithms and underlying data may be the focus of much scrutiny, and questions on ethics and human rights come to the fore — what’s been largely missing is a deeper understanding of why the problems that these technologies seek to address are occurring, and how these ‘automated solutions’ actually affect the resilience and performance of these industries.

Algorithmic transparency alone cannot comprehensively examine the inter-relationships within these contexts, or how they are changing as a result of these technologies. Why? Because historical data and a rules-based ethical framework cannot accommodate a continually evolving world, especially when not everything can be measured, and much in our world is ‘trans-contextual’ and learning or responding to the changing conditions around us.

At the heart of the matter lies many questions of how these technologies shape our society, what powers of control we have in directing these changes and how we perceive and think about fairness. Do we really want to go down the path we’ve prescribed for ourselves?

Good, Better, Best…

“We can be fully human without being in complete control of our world” — Douglas Rushkoff, Team Human

The business of decision-making is fraught with judgement — choosing between various alternatives, discriminating between different features and weighing up multiple possibilities — are all in the hope that any actions taken as a result of these decisions will achieve the desired outcome.

Underlying this human desire to predict and therefore ‘control’ outcomes to a pre-determined future lies the implicit assumption that these decisions will shift the system (be it an organisation or a society) to a ‘better’ state.

These ideas are not new — one historical example being the desire for ‘improving humanity’, which gave birth to “Eugenics”(improving the genetic composition of the human race). First originating during the time of Plato (around 400 BC) and later developed after being inspired by Darwinism in the early 1900s, eugenics is supposed to literally mean “good creation”.

Contrary to popular belief, Darwinism does not fully explain the phenomenon of evolution— Mendel’s research into the hereditary and variation in peas, along with William Bateson’s interpretation of Mendelian principles, suggest that it could not explain ‘new species’. Interestingly, at around the same time, English Philosopher, G. E. Moore, in his Principia Ethica (1903) contended that “good” could not be defined.

These insights are important when using AI technologies in ‘selecting’ features or, in the example of recruitment and employment, choosing humans for a particular job or task. What this means is that what is ‘good’ and what is ‘fair’, is not only open to interpretation — but that even if these could be agreed on, the outcome that’s been engineered may not be as robust as we may have thought!

In his book, “Out of Control”, Kevin Kelly talks about how “a little touch of randomness… actually creates long term stability”. So what might appear to be sub-optimal choices, could actually be critical elements in ensuring the resilience of systems through diversification! Moreover, Kelly emphasises the importance of ‘symbiosis’ (mutually beneficial interactions) in relationships, noting that in “one mutual relationship, evolution could jump past a million years of individual trial and error.”

What AI applications are as yet unable to capture are elements of ‘mutual learning’ or ‘symmathesy’, as Nora Bateson refers to, which are dependent on the many contexts and the responses within that environment. It is within these ‘learnings’ that evolution and adaptation is occurring.

In our earlier example of AI recruitment, determining the optimal facial features, which in turn supposedly pre-determines personality traits, misses the inter-relationships and learnings that happen within an organisation. Not only that, the opportunities for innovation and growth within the organisational ecosystem are also limited.

AI technologies may be able to optimise for what we think are ‘best case scenarios’, but may be dismissing key attributes and features that are essential for our long term viability — whether it be organisation, industry or nation.

Phoensight is an international consultancy dedicated to supporting the interrelationships between people, public policy and technology, and is accredited by the International Bateson Institute to conduct Warm Data Labs.

AI, AI on the wall —  Who’s the Fairest of them all? was originally published in Phoensight on Medium, where people are continuing the conversation by highlighting and responding to this story.

I was nine when I really discovered that we weren’t living on a Flat Earth — it was when an orange I held in my hand didn’t conform to Euclidean Space. In school I was taught that the angles in a triangle summed to 180 degrees, but I could see that the triangle I’d drawn on my orange had at least two right angles!

Fast forward a few decades, and I’m still in awe when nature refuses to fit into the models and paradigms we create. Not just oranges, but people too. I continue to be fascinated by the differences between the outcomes predicted by ‘data-driven’ models and the way people respond. And for a long time I’d apportioned much of the blame for this discrepancy on the limitations of these models and the data used.

And while I still believe this to be true, it took a shift in my perception of these ‘limitations’ to uncover some new insights. You see, the assumption of a flat surface is critical for the angles in a triangle to sum to a straight angle — but the many varied ways and contexts in which this assumption fails is where the interesting phenomenon and insights lie undiscovered.

Outliers on the Inside

A longstanding student of data modelling, I’d fallen into the trap of swapping ‘reality’ and ‘abstraction’ — and the ‘outliers’ in my data, apart from causing me much angst, were rarely the main game. Often omitted and side-stepped, these were put into the “difficult” basket with the understanding that if there was an underlying reason for the deviation, it would require a different model, separate study or at least an alternative hypothesis to be tested.

At the same time, there was always an uneasiness about ignoring outliers — “don’t throw out the baby with the bath water!”, was what I got told by more seasoned data professionals. “You might be dismissing valuable information which could be the missing piece of the puzzle.” I learnt that data scientists had developed many methods in their toolkit for how outliers may be dealt with.

But what was missing in my early attempts in coaxing these outliers back to the ‘modelling fold’ was a deeper understanding of the various contexts that this data was being extracted from.

Instead of my models truly describing reality, they were ‘fitted’ to the data. When describing the physical sciences, this worked remarkably well — for living systems, however, my results were mostly inconclusive. So I struggled to find models that really suited the data, and nature found it difficult to fit into my human-made abstractions!

But still I persisted — fine tuning assumptions, including higher order terms and fitting different types of curves to the data. Yet, at no point during those data cleansing and modelling processes did it occur to me that some of these ‘data critters’ were part of a tangle of contexts that held information of a different sort — information that might be difficult to capture through recorded data. Information that refused to be pickled in a database!

What I’d often put down to ‘noise’ in a stochastic process, was still objective in my mind — random, but objective. And any information that could be categorised as ‘subjective’ was still being processed using the same ‘data pipelines’. To me, ‘subjective data’ was information that was somehow biased because it came from particular perspectives.

And describing nature was one of those messy things — the nearest I got to that was via qualitative information. Still, nature refused to stay true to that description, and kept on moving and changing…

Nature’s Alchemy of Objective and Subjective Data

The novelty of being able to reproduce scientific results was what first made me love the scientific method — rediscovering many years after the fact that the acceleration due to gravity was the same, no matter how heavy the object in free-fall, was like a parlour trick! There’s something about being able to ‘objectively’ verify results and conclusions that give it credence.

Experiment, change a variable — holding everything else constant, and then experiment again…

The ability to collectively predict an outcome using the same methodology comes with its own confidence — and this also allows others to build on your work. There’s a certain rigour and transparency that allows for greater agreement on decision-making with ‘objective’ information. And ‘objectivity’ lends itself well when ‘scaling’ solutions. So far, all’s well — until it’s not!

Living systems cannot be ‘predicted’ so easily. At many points along my journey I blamed my ‘imperfect datasets’, with a range of complaints: small/sparse sample sizes; self-selection biases; missing information or inaccuracies — thinking that if I had the perfect dataset my models would describe reality perfectly. Looking back, I had discounted the many contexts that the data from living systems had come from.

Why? Because, for the most part, in an ‘objective world’ the context of these solutions doesn’t play a significant role. Subjective data, however, is different — context is everything! Context creates the conditions for how we percieve and respond in our world. Not only that, but living systems are continually learning and adapting through numerous interactions in their environment.

When the Greek philosopher, Heraclitus, observed that, “You could not step twice into the same river,” he was alluding to this elusive, continually changing quality of nature and living systems. One that won’t allow for Ceteris Paribus (holding everything else constant).

My inability to work with continually shifting subjective information and my confidence in working with ‘objective’ data had me subconsciously prefer the later. Yet, subjective data was no less important — and thinking that qualitative data was the same thing was missing the visceral nature of this information... Nature seemed to be home to both kinds of information.

My data science journey with objective and subjective data was at a dead end— that is, until I discovered Nora Bateson’s “Warm Data Labs”. Nora describes “warm data” as information within the inter-relationships that’s ‘interwoven in complex systems’.

“Warm Data is not meant to replace or in any way diminish other data, but rather it is meant to keep data of certain sorts “warm” — with a nest of relations intact.”

— Nora Bateson

This continually changing subjective information may not be able to be collected, but working with this information — by understanding the many contexts and dependencies allows for new insights and also maintains the integrity of this ‘warm data’.

By giving subjective data the respect it deserves, and allowing it to flow freely outside our human-created models it might be a willing partner to what we call ‘objective’ data. But even then the distinction can become blurry — how do we really know which is which? To do that we need to be vigilant and look for what’s hidden within a forest of changing contexts…

Phoensight is an international consultancy dedicated to supporting the interrelationships between people, public policy and technology, and is accredited by the International Bateson Institute to conduct Warm Data Labs.

What Nature taught me about Data Science was originally published in Phoensight on Medium, where people are continuing the conversation by highlighting and responding to this story.

At a time when misinformation runs rife in social media and while governments are grappling with keeping up with controversial public discussions, how we allow for complexity will play an important role in progressing forwards.

Every policymaker’s manual has some fine print on ‘complexity’ — and if you’re lucky, you might find a chapter or two about it. There’s an unspoken reluctance to open the Pandora’s box on complexity to the wider public. Why? There may be a multitude of reasons on the checklist, but you’ll find ‘explainability’ featuring strongly in there somewhere — like a prickly cactus no one wants to touch.

The nature of complexity means that it’s difficult to end up at a pre-determined outcome by following a prescribed path — assuming that outcome even solves the problem you’re working with. Yet, in a world where countries aspire to provide more transparency, complexity is often side-stepped or hidden behind a series of tweets, political slogans, announcements, action plans or reduced to tables, charts, dashboards and fancy infographics — or worse, cited as a reason for a lack of transparency!

Now, don’t get me wrong — there have been many successful action plans, and visual representations have been invaluable in helping people better understand and explore information. But if we take a step back to the issues of concern, there are many questions that need to be asked at the outset — “How have the issues been framed?”, “How do we determine what information is relevant?” and “How does the way we scope a problem affect the solution?” amongst others.

And this is where the ‘taming’ of the complexity is actually taking place…

Framing the Problem is putting it in a box

The complexity within which an issue is held — in contrast to the complexity within the issue— is ‘reduced’ through the articulation of the problem itself. And what we often perceive as the ‘problem’ is a representation that’s derived from this reductionist approach.

Some will recognise this as seeing the problem through the lenses of various political propaganda, social media filters or subconscious biases. There is the potential for an enormous amount of variability in even defining the problem.

So the process through which a policy problem is articulated is where the transparency ought to begin! Before the problem is framed… with observations of and across the many contexts around the issue.

Policy modellers may recognise this as having to explain the rationale as to how they came up with their models in the first place, what they perceived as the problem, their assumptions and modelling limitations — not just releasing their models as open source, which in itself is a step forward in the Open Government movement.

Daunting? Yes, but this may be because the original ‘complexity ecosystem’ is no longer present to support the many contexts that have been stripped away while articulating the problem. That, and explaining rational decisions in complex living systems doesn’t come without a hitch… sooner or later you’re going to have to confess using that secret ingredient — your gut feeling!

And this takes us back to the source of the complexity…

Can’t see the Forest… without looking at the Trees

“The major problems in the world are the result of the difference between how nature works and the way people think” — Gregory Bateson

There have been a lot of discussions on how complexity should be ‘managed’ or ‘reduced’ — and many a policymaker has been challenged with delivering a ‘simple’ message or solution for the masses to ‘understand’. Indeed the study of ‘complexity’ itself and the many steps, tools and frameworks to ‘tame’ it can be intimidating enough for most people.

Then there’s the language around it — complexity vs complicated (I’m particularly like Mark Foden’s metaphor of cats for understanding ‘complexity’ and clocks for ‘complicated’, but it gets trickier when you notice that some cats act like clockwork at mealtimes…!), linear vs non-linear or lateral thinking, static vs dynamic, stable vs chaotic…

Can you picture a diagram of boxes and circles on a whiteboard, with arrows pointing outwards and inwards in at least half a dozen colours? Or circular flowcharts describing processes? While these may be useful abstractions in ‘modelling’ complexity, they are still a step away from the complexity itself!

So how is transparency ever to be achieved amidst all this? Is a reductionist approach that aims to define a problem at the outset inevitable?

At this point we find ourselves at the perceived dichotomy between complexity and transparency…

In a previous article, we discussed the relationship between ‘transparency’ and ‘openness’ — and how information may be ‘open’, but without ‘sense-making’ there can be little transparency or understanding.

While the purpose of ‘Open Government’ is to allow for citizens to have some public oversight into the workings of government, the sense-making involved in this process is critical for genuine transparency.

This means that any interpretation or articulation of the issues might actually distort or even obscure the true nature of the complexities involved rather than serve to simplify them. So where does a policy maker start when looking at issues within society?

There is a distinction to be made between ‘reducing the complexity by simplifying a problem so that it may be easily understood’ and ‘creating the conditions for allowing a deeper understanding of the complexity’. One implicitly re-frames the complexity for public consumption through a pre-determined lens, while the latter allows for the public to re-frame the complexity using their own sense-making faculties directly from the source.

While there are many overlapping moral and ethical contexts involved in comparing the two, this takes us into the hazy space of ‘trust’ in government— and the articulation of public and social issues. But it also touches on the perception of the capacity of the general public to understand or be willing to work with the raw complexity.

But just as in nature, people are working within and with complexity all the time …

Pure, Unadulterated Complexity

“We can liberate ourselves by trusting our own instinct and finding the thought-lenses which show us our world in the way we need to see it…” — John O’Donohue

In continuing the work of three generations in her family, Nora Bateson and the International Bateson Institute have carefully created group processes known as ‘Warm Data Labs’ that allow people to develop a deeper understanding of complexity through multiple contexts.

Warm data labs create the conditions required for people to work with complexity through ‘trans-contextual’ lenses — be it economic, educational, cultural, ecological, health or technology, amongst others — so that they are able to develop a deeper understanding of the interdependencies within and across these many contexts.

The embodied mutual learnings that occur in warm data labs allow for new patterns and insights to be identified. In doing so, they also deepen the understanding of the issues themselves, so that new societal responses may be possible.

So here is one example where carefully allowing for complexity actually allows people to find new pathways for sense-making in ways that suit their individual needs. And far from increasing the barriers to understanding, the way in which complexity is worked with provides more avenues for sense-making — and therefore a deeper comprehension of the underlying issues. Better transparency into public policy and open government might be possible on a whole new level!

Considering the issues with a deeper, more nuanced understanding might allow for better and more effective solutions at a systems level. Of course, publicly communicating and implementing these solutions would also be susceptible to the same reductionist traps we discussed earlier!

Perhaps there’s a reason why the origin of the word “complex” stems from the latin word, “complectere”, which means embrace…

Phoensight is an international consultancy dedicated to supporting the interrelationships between people, public policy and technology, and is accredited by the International Bateson Institute to conduct Warm Data Labs.

Why Complexity should be Embraced for Transparency was originally published in Phoensight on Medium, where people are continuing the conversation by highlighting and responding to this story.

When Google’s Chief Economist, Hal Varian, said that “Google will make us more informed,” he was actually referring to how greater access to information may increase people’s potential to contribute to society. And while this may be true, digital technologies are also subversively changing the interrelationships between governments and businesses.

While organisations are continuing to digitally transform, everyday transactions are increasingly being mediated by technology and captured through data. But it’s much more sophisticated than merely translating what once used to be analogue information into digital — technology (including smart sensors) has the ability to collect and record much more data on a single transaction than humans have ever done before.

Not only that, the computational processing power and storage capacity that organisations have access to today are phenomenal. Compare this to the early 19th century, when censuses were carried out by collecting paper documents and collated by hand. Data assets were historically held and predominantly managed by Governments, and with that came their ‘monopoly’ on citizen data.

Our ‘citizen data assets’ saw governments organising their efforts into various portfolios such as finance, health, education and social services (amongst others), and the collection of civic data became more sophisticated and aligned with departmental services. Policy advice and economic forecasts also benefited from more data — with government departments in the best position to provide any quantitative ‘crystal-balling’ of the future.

Much has changed today and — like it or not — private organisations such as Google, Facebook and Amazon are now holding the lion’s share of information on citizens. The large masses of data they collect gives them unprecedented access to personal, social and economic information, allowing them to derive new insights and potentially achieve forecasting supremacy (if they haven’t already!).

Meanwhile, governments are still grappling with the many challenges of data regulation and integration — and are coming to terms with how to get the most out of their data. Government data, often fragmented by the very portfolios they put in place to compartmentalise their services and policy advice, has become ‘siloed’. And this has also constrained Government’s ability to compete with the private sector for deeper data insights.

The balance of power when it comes to knowledge about citizens has undoubtedly shifted from Government to Business.

Yet, the narrative is so much more complicated — governments have a duty of care for their citizens, and this includes protecting citizen privacy and human rights. The apparent tension between deeper data insights and privacy protection has been the focus of many government data integration projects in an effort to gain more ground in the understanding of complex policy issues.

Policymakers today have to navigate not just government agency silos, but a public-private collective of information that could be utilised to create better public policy. Private data collected in organisations is mostly used to uncover new business insights and better customer/client services. But this same data, viewed from a social good perspective, could also provide new policy insights and benefits to society.

As data sources continue to diversify, we’re beginning to see new data-sharing collaboratives emerge, with McKinsey & Co. estimating an increase of around 20 per cent in companies forming data-related partnerships from 2017 to 2019. By collectively putting the pieces of the ‘data-puzzle’ together, better data models are being developed for better services all round.

Indeed, the recent COVID-19 pandemic has resulted in unprecedented data-sharing arrangements which has enabled faster outbreak research. At the same time, Google has been using location data gathered from smartphones to help public health officials understand people’s movements and compliance to new policies.

The task of data sharing, though noble, does not come without its challenges — data sharing collaboratives present a myriad of legal, technical, commercial, ethical and privacy risks, not to mention its impact on public trust. What’s required at the outset are strong governance practices that ensure responsible data collaboration and innovative solutions.

Last month, the ‘Beeck Center for Social Impact + Innovation’, supported by the Rockefeller Foundation, released a guidebook on sharing data for social impact, which includes a framework for establishing responsible governance practices. It acknowledges a ‘fundamental need to reform data-governance’ in an environment that is often under-resourced, leading to practices that can be unsustainable and ineffective.

And this leads us to a closer examination of the relationship between governments and businesses.

Interestingly, while government may have lost some of it’s citizen data ‘turf’ to the private sector, it still has the power to protect the interests of its citizens and address any imbalance through carefully considered rules and legislation. The ‘General Data Protection Regulation (GDPR)’ adopted by the European Union is one such example where legislation is redefining how businesses uses personal data.

But can a fundamental change of the relationship between government and business, as it relates to data, achieve a greater social good?

In her recent Schumpeter lecture, Professor Mariana Mazzucato, founding director of the ‘Institute for Innovation & Public Purpose’, spoke of the need to re-design the governance systems of public and private sectors to redress imbalances within capital markets, including how they relate to one another. By focussing on the concept of “value”, Professor Mazzucato puts forward what might be an evolution of public-private partnerships — ones that are ‘mission oriented’ to co-create greater public value.

Could the same may be applied to data-driven or evidence based public policy design? The prolific growth and re-distribution of citizen data assets may warrant a re-think of the role of government and business.

In 2018, the Australian Treasury and LinkedIn presented some early findings on the gender gap in tech entrepreneurship across seven countries. What was unique to this pilot, was that almost no data was exchanged between parties — what was shared were aggregate data insights that were relevant to the policy problem. Here was a case study that showcased what might be possible with collaborative data efforts focussed on a specific ‘mission’ and a small shift in the role of government and business.

How the relationship between government and business will evolve with more data is as yet unknown. But what is clear is that technology has changed the status quo, and the role of government and business must shift if we are to achieve greater social value with data.

— — —

Phoensight is an international consultancy dedicated to supporting the interrelationships between people, public policy and technology, and is accredited by the International Bateson Institute to conduct Warm Data Labs.

This article was based on ‘Who’s running things now? Why data is changing the roles of government and business’ Mandarin, Feb 2020, and the ‘Data for Good’ Event, hosted by General Assembly on the 5th Feb 2020.

Technology is changing the role of Government and Business was originally published in Phoensight on Medium, where people are continuing the conversation by highlighting and responding to this story.

As new measures are ushered in to help our economic, financial, social, education and health systems cope with a pandemic, the interdependencies between these systems are becoming more obvious. Much less conspicuous — but still important — is the impact of Artificial Intelligence (AI) on society and these very same systems.

Todays emerging technologies bring to the fore many questions relating to our society, our collective values and how we interact with one another and our many systems. While the challenges of adapting to an ever evolving ‘technological ecosystem’ are complex, the many issues and unintended consequences that surface are interdependent in ways that may not be easily perceived or understood.

Public concerns about algorithmic bias and fairness, data privacy and various forms of digital divides, such as accessibility, take up of AI services and inequality, put a new focus on the vulnerabilities in our existing systems. This is not just about ‘fixing’ AI— it’s about understanding the many complexities within our social, economic, political, regulatory, educational, health and cultural frameworks — and the moral or ethical undertones within these.

The ethical implications and social impacts of AI have been the focus of much research and debate — and with these come a new wave of approaches and tools, such as the AI Fairness 360 Open Source Toolkit, to try to address some of these ‘technological shortcomings’. But so far, there doesn’t appear to be an easy solution to the problem.

There is a tendency to either address the symptoms that AI surfaces or embed measures of fairness and ethical considerations in the design of AI algorithms. While these go some way towards solving the problem, neither seek to address the deeper issues prevalent in our society that traverse multiple contexts.

What is neglected is a closer examination of the many interconnected systems we are a part of — and how these are evolving with and in the presence of emerging technologies. What we are witnessing, as a result of the unintended consequences of AI, are the distortions within our existing systems.

What’s more is that the presence of new agents, such as AI (or even a pandemic), are testing the integrity of our systems — and how they adapt to a changing ecosystem. What we learn from this, and how AI shapes our society, will determine how we design and build new technologies — but it also informs us about the design and limitations of our systems.

How are we adapting?

Building AIs that make decisions with very real human consequences, is taking us into uncharted waters that are more nuanced, and require a closer examination of the interrelationships and interdependencies within our complex societies. New generations of AIs provide us with new avenues and contexts to understand how we relate to each other — and at the same time are transforming our values, preferences and behaviours.

We cannot separate the influences that AI has on our society from the influences that society has on AI — the two are inextricable. Developing AI regulation without this understanding, in an attempt to ‘moderate’ AI, may result in other unintended consequences for society.

What’s needed is to work within the complexities of the ‘AI-human ecosystem’ — and this means understanding the many interactions, interdependencies and interrelationships within our changing ecology. It also means closely examining how our natural and human-designed systems relate to each other across many contexts. By doing so, we may reflect on the suitability of our other human-designed systems in their currently evolving state.

For example, recruiting technology firms using AI to analyse facial movements to determine employment suitability for certain sectors will undoubtedly alter those industries themselves. The data and algorithms used may leave an indelible imprint on the recruitment process and the sectors that use it — and this may be partly characterised by certain biases.

The question we must ask ourselves is whether we understand all the characteristics of these ‘imprints’ — and whether or not these are desirable or intended in our society.

A deeper question around the patterning of an AI algorithm would be — how do we identify these characteristics? The entangled nature of our technological and human complexities mean that these imprints or characteristics may surface in other contexts. And so these may hold clues as to where we need to look…

In the example of AI facial recruitment technology, whilst regulating the use of AI may address some of the social issues, a closer examination of the contexts around health, economics and finance may reveal new ways of finding suitable employment candidates — which result in better outcomes. It may be possible that re-designing the journey to employment could ‘side-step’ the need for AI being implemented in that manner.

Warm Data Labs

What’s been missing from the AI narrative so far, is a broader approach — one that encompasses a rich and diverse ecology of the many evolving contexts that are tangled together, thereby creating countless stimuli for mutual learnings.

These mutual learnings or ‘Symmathesy’, as described by Nora Bateson, founder of the International Bateson Institute (IBI), go deeper than just taking a ‘systems perspective’ — because they also include the ‘expression and communications’ of the interdependencies within living systems. By working within these complexities we may discover new avenues and ways of sense making for a deeper understanding of our systems. And it is in these learnings that we may be able to identify the patterns of discordance within AI.

But where does one begin with such an approach?

Pioneered by the IBI, ‘Warm Data’ is the continually evolving information that’s held within the interactions of living systems across many contexts. ‘Warm Data Labs (WDL)’ are carefully designed group processes that help provide the conditions to explore in complex systems. These labs allow participants to become ‘immersed’ in the complexities they are working with to get a better sense of the underlying inter-relationships, and to identify new insights or systemic patterns that emerge across contexts.

As the nature and attributes of technology are unprecedented in their speed, reach and impact, any policy instrument or regulation that does not consider the broader contexts, may also result in unprecedented unwanted consequences. Warm Data Labs may allow for a deeper understanding of the consequences of new technologies such as AI, as well as the effects of any actions, policies or regulations within our changing systems.

Any proposed strategy, policy or research on AI regulation will need to consider the impacts across and between multiple contexts. This includes a re-examination of issues relating to social justice, human rights, consumer protection, ethics, data privacy and protection, surveillance and transparency within and across our existing systems — amidst new technologies.

Designing Systems for Fragility and Vulnerability

Humans have evolved through many complexities that defy logic in favour of protecting the fragile and vulnerable. When Margaret Mead, an American anthropologist, was once asked what she considered to be the first evidence of civilisation, she answered — a human thigh bone found in a 15,000 year old archaeological site that was ‘previously’ fractured, pointing out that for a person to survive a broken femur required them to have been cared for long enough to be healed.

It may not be possible to understand human behaviour, emotions and the reasons for them (in their multifaceted complexities) by decontexualising this information into collected data. The complexities that are held within the human psyche are formed within multilayered contexts and numerous life experiences. But this may hint at how our systems could be designed.

In his essay on the evolution of humanity, French geophysicist, Xavier Le Pichon puts forward the idea that, like with plate tectonics, vulnerabilities within a system are important for facilitating its evolution. He notes that a system that is ‘too perfect’ is rigid and closed, and can only change through great disruption.

Le Pichon postulates that, contrary to what’s often assumed, tending to system vulnerability is required for evolution, and any attempt to overlook this may prevent it from functioning at all. He argues that the evolution of human society through the years has benefited from ensuring that the fragile and vulnerable in communities are protected.

Perhaps the best designs for AI and technology will focus on tending to the vulnerable (across many contexts) — and go beyond just inclusion. The same might be said for our other systems. Designing systems for fragility may help ensure that the many interdependencies within our ecosystem are better regulated.

Understanding the interrelationships within our many systems may enable them to be re-designed to uphold the values of our changing society in the presence of AI — instead of conforming to it.

Phoensight is an international consultancy dedicated to supporting the interrelationships between people, public policy and technology, and is accredited by the International Bateson Institute to conduct Warm Data Labs.

Designing AI Regulation is really an opportunity to re-design our Systems was originally published in Phoensight on Medium, where people are continuing the conversation by highlighting and responding to this story.

As Artificial Intelligence (AI) permeates deeper into our society, new manifestations of perverse societal outcomes are being unleashed. Public concerns about algorithmic bias and fairness, data privacy and various forms of digital divides, put a spotlight on the existing cracks in our systems.

The ethical implications and social impacts of AI are the focus of much research and debate — and with these come a new wave of approaches and tools to try to address these ‘technological shortcomings’. But, as yet, there doesn’t appear to be an easy ‘patch’ to the problem...

We are beginning to see the distortions in our systems through an AI lens. What we learn from these, and how AI is shaping our ecosystem, will determine how we design and build new technologies — but it also informs how we design and understand our systems.

What we are coming face to face with is a disruption of our systems as we know them to be — spurred on by the machines we have trained. This is not about an ‘AI takeover’ — it is about understanding the complexities of our social, economic, political, regulatory and cultural frameworks, and the moral or ethical undertones that are ‘baked’ into these.

Building AIs that make decisions with very real human consequences, is taking us into uncharted waters that are more nuanced and require a closer examination of the interrelationships and interdependencies within our complex societies. New generations of AIs provide us with new avenues and contexts to understand how we relate to each other.

At the same time, the creation of these new forms of intelligences does not come without its responsibilities — more so when we train technologies to teach themselves on data and processes that inherently capture the best and worst of humanity. Without them having the ability to contextualise their learnings or comprehend the broader implications of their decisions as they evolve, the use of AIs might drive humanity towards unimagined consequences.

We cannot separate the influences that AI has on our society from the influences that society has on AI — the two are inextricably diffused. Bolting on regulation, without a systems perspective, in an attempt to hold AI back from running wild may have unintended consequences for society — as will allowing it to run free, untamed and unchecked.

What’s needed is to work within the complexities of the AI-human ecosystem — and this means understanding the interactions, interdependencies and interrelationships within this changing ecology. It also means closely examining the differences between AI and human intelligence.

This is no easy endeavour. For one, AI’s tend to exhibit a ‘black box nature’, and are known to sometimes surface unexpected outcomes — and for another, we need to be comfortable and familiar with the many complexities in living systems and our societies. So where does one start with something so daunting?

Well, as it happens, humans have encountered something just as enigmatic and difficult to explain through the ages that may hold some new insights— love.

Love in Binary

“When the knot is hacked in two, a binary is born and the poem is lost — not solved, but truncated.”

— Nora Bateson, from ‘Small Arcs of Larger Cirlces’

In ‘The Matrix Revolutions’ when the Oracle tells Sati that “cookies need love like everything does”, she hinted at what AI is missing — love is the quintessential human emotion that has fascinated poets, artists and philosophers through the ages. Subjective by nature, difficult to explain and sometimes unpredictable, many have attempted to express love’s elusive qualities through the creative arts with varied success.

AI researchers and developers find themselves in a similar conundrum — with many on a quest to breathe empathy into their AI creations, in the hope of improving AI performance and decision-making. That human relatability to machines could be supported by emotive technology is seeing new advances in AI — but the challenges might be as old as the mythical ‘Golem’ developing a cruder version of human emotion, only to eventually cause havoc.

What we are dealing with is much more complex than simply coding in ‘empathy rules’, based on collected human data, to upgrade our AI algorithms. Besides, mapping out the effects of human emotion says little about their causation — let alone the processes involved in their evolution. There is a difference between developing AI to emulate human emotions and designing AI to enhance and support our ‘human-ness’ within its messy complexities.

The ocean between how we perceive the role of emotions in our lives and their true origin may be as deep and complex as sifting through innumerable life experiences and across multiple contexts — ecological, cultural, and economic amongst others.

Studies identifying many interdependencies between our cognitive, emotional and social intelligences within our neural systems suggest that our ecology and environment contributes significantly to our human intelligences. While the phrase, “Beauty is in the eye of the beholder”, has been held in greek literature since the third century BC, science has confirmed that an individual’s aesthetic preferences are mostly based on personal experiences (not genes!) — which are highly subjective.

The ecology matters — and teaching AI about human emotions through collected data is like translating the anguish of waiting to see a loved one through the number of days that pass, photos and messages… What’s missing is the information within a loving gaze, which cannot be interpolated between data points. It can only be learnt through the complexities of life experience by interacting within the many and varied complexities of our ecosystems.

Love continues to defy logic — and evade technology’s attempts to ‘bottle’ its essence in a data warehouse or algorithm. No algorithm can be trusted to blend the infinite complexities held deep within each human psyche — especially if they are continually shifting!

Explainable AI may be able to identify which features were responsible for your credit rating, but might miss that your preference for sunflowers comes from feeling a sense of freedom and joy— and that one story at the seed of those emotions comes from a memory you still hold tenderly of planting sunflowers in the garden with your beloved Grandmother.

Not all complexities are created the same— the complexities of systems created by humans, such as our financial, political and economic systems are different to those that exist within nature. And so it follows that what we learn from interacting with different types of complexities must also be different. The complexities that AIs work with, while processing large amounts of data, are different to ones we encounter in our lives. The same might be said for how each encounters love…

And as to why the sciences were unable to snare love into logic, and why the arts could only trace its outlines? Perhaps it has something to do with the balance and interdependencies of both, objectivity and subjectivity. While logic holds firm its door to entry, the arts are much more permissive allowing more entries to experience its offerings through different perspectives and contexts.

Yet, both cannot really capture this ever evolving quality we call love. Just as no amount of poetry can compare to the experience of falling in love, no amount of data or algorithmic programming can be the same.

Love is analogue, not digital.

Context is (a subset of) Everything

“I read in a book once that a rose by any other name would smell as sweet, but I’ve never been able to believe it. I don’t believe a rose would be as nice if it was called a thistle or a skunk cabbage.”

― L.M. Montgomery, from ‘Anne of Green Gabels’

AI is not the first to be perplexed by how humans behave and make decisions. Another came before — Economics. For a large part of its evolution, the study of Economics focussed a great deal on the ‘rationality’ of humans, who would ‘consistently apply their preferences’ — until it became apparent that humans did not behave in line with their economic model counterparts!

Economics, dissatisfied by its attempts to accurately predict and influence market outcomes, incorporated insights from psychology, neuroscience and microeconomic theory to better understand and shape human responses — and ‘Behavioural Economics’ was born.

We are witnessing a similar story with AI. With increasing computational processing power and data storage capacity, AI is evolving by incorporating new datasets, such as those used for facial recognition and emotional recognition, to better predict and respond to human behaviours and preferences.

While the ultimate goal of AI is widely touted to be reaching ‘singularity’ (i.e. when AI surpasses human intelligence), what’s still not clear is how these will be compared. In his recent lecture at Lafayette College, Pennsylvania, Kevin Kelly spoke of the multiple dimensions and taxonomies of intelligence — some of which are yet to be identified, and some that might be unique to animals and other species. And these are constantly evolving!

Understanding human behaviour, emotions and the reasons for them (in their multifaceted complexities) may not need untangling and dissecting. Perhaps like love, formed within multilayered contexts and numerous life experiences, objectivity and subjectivity, rationality and irrationality, science and art must be blended together for a new type of sense-making.

What’s been missing from AI so far, is a broader approach — a rich and diverse ecology of the many evolving contexts that are tangled together, which create the countless stimuli for mutual learnings. These mutual learnings or ‘symmathesy’, as described by Nora Bateson, founder of the International Bateson Institute (IBI), go deeper than just taking a ‘systems perspective’ — because it also includes the expression and communications of the interdependencies within and across living systems.

Perhaps, by working within these complexities (rather than reducing them, which limits our understanding) we may discover new avenues and ways of sense making — ones that allow for a trans-contextual exploration of our complex relationship with ‘intelligence’. We need to ask ourselves what we mean by ‘intelligence’ in a changing world, by immersing ourselves through a kaleidoscope of contexts, so that ‘intelligence’ may be allowed to shift and change — and adapt to the many systems we’ve fenced around it.

But nothing is really so disconnected from everything else, and a shift in our perception of intelligence (artificial or otherwise) will ultimately also shift the way we perceive and design our many systems. ‘Warm data labs’, developed by the IBI, provide the conditions to do just that — and may prove to be an invaluable process to achieve this.

One more thing about love — though St. Clare of Assisi may not have been thinking of AI when she wrote, “We become what we love and who we love shapes what we become”, she must have known that how and to whom we respond to plays a part in our evolution.

How we perceive and relate to any intelligence, including within our existing inter-species relationships, will undoubtedly contribute to our own ecological learnings — which then informs how we design both AI and our other systems. The same might be said for how we relate to AI — will responding to other intelligences in a way that’s apathetic change us differently to if we responded with kindness or love? If so, how do we create the conditions that allow for more trans-contextual complexity in our relationship with intelligence?

Danielle Krettek (founder of Google’s Empathy Lab), drawing inspiration from Jane Goodall amongst others, has shifted the AI design narrative from ‘design thinking’ to ‘design feeling’ — emphasising the importance of our social and emotional connections. Her research shows that when technology is ‘attuned’ to strengthen our emotional connections, an ‘empathic leap’ is possible with machines.

We have the chance this time to go further — we have the unprecedented opportunity to re-design not just AI, but also our human created systems using a trans-contextual approach… where we can work within moving complexities and allow for new shifts in our perceptions.

Perhaps AI needs love — like everything does.

“Carl Sagan and Ann Druyan taught me a lesson that I now hold as my secret mission: when designing for AI, be sure to sneak the love in” — Danielle Krettek

Phoensight is dedicated to supporting the interrelationships between people, public policy and technology, and is accredited by the International Bateson Institute to conduct Warm Data Labs.

AI needs love like everything does was originally published in Phoensight on Medium, where people are continuing the conversation by highlighting and responding to this story.

The surging digital transformations that organisations are experiencing are a result of being swept through a strong digital current. Companies are racing into big data technologies, embracing tools like AI, whilst navigating through the tricky terrain of very real human issues such as data privacy and ethics. The liminal interface where humans meet technology is where we’re seeing the rubber meet the road.

Digital technology is now well and truly a part of our ecosystem, and to perceive it as merely a ‘tool’ in the story of our human evolution is to miss the many ways it’s shaping and shifting our society. If anything, our relationship with data and its excessiveness has brought into focus many questions on what role humans are playing in this evolutionary process — and what conditions are required to achieve the best possible outcomes.

We are still grappling with how big data and the information humans carry are intermingled — how one is influenced by the other, and what the resulting ‘information melting pot’ means for organisations and society. This may sound like some form of ‘digital alchemy’, but to really benefit from big data we need to examine these inter-relationships closely and observe how they are are evolving.

Of course, this is no easy task as humans hold vast amounts of information gathered during their lives through innumerable interactions with (and learnings from) their environment. These may be personal, subjective or coloured through cultural, social and economic perspectives across many contexts!

Yet, this information is invaluable in how we make sense of the world around us — and warrants closer study.

But how do we go about unlocking it?

The International Bateson Institute (IBI) has recently introduced the concept of ‘Warm data’, which is information about the interrelationships and interdependencies within complex systems across many different contexts. ‘Warm data labs’ are group processes where insights into warm data may be gleaned by working with information through ‘trans-contextual’ lenses.

The ‘magic’ in this process is that considering a complex issue across several contexts allows participants to become immersed in and have a better understanding of the complexities involved. Not only that, these processes allow for changes in perspectives as people explore the interdependencies within their multiple experiences.

Here is information that is not just unlocked, but evolving in real-time with new learnings — this is human agility!

What’s exciting about this, from a big data perspective, is the opportunity to explore the potential of warm data and how it may inform big data strategies for better outcomes — by providing the right conditions required to unlock valuable human information.

Let’s briefly look at two applications: digital transformation in organisations and data science.

Digital Transformation

Using big data for delivering better services is seeing many organisations undergo digital transformations for market survival, with worldwide spending expected to reach $2 trillion by 2020. What’s surprising in all this is that many companies receive poor returns on their investments, with Forbes estimating a waste of $900 billion of the $1.3 trillion spent on digital transformation in 2018 alone.

The challenges that companies face during such transitions go beyond just the technology adopted, and include supporting human capability, organisational culture and investment decisions etc. These are no strangers to any CDO, and many organisations are focusing on data strategies that aim to empower people with the right information to facilitate organisational transformation.

But where is the right information held?

Is it in the big data we’re collecting en masse? Or is it held within an organisation’s workforce? Or is it a mixture of these?

The warm data or the trans-contextual information at play here is dynamically changing and evolving — and within an organisation is held across the business silos, the experiences of the people working there, the interactions with stakeholders, the customer relationships, the judgements and assumptions made by data modellers etc.

While we’re figuring out how to improve the synergies between big data and the information that people within an organisation hold, recognising that we need a deeper understanding of the ecosystem across many contexts is critical to delivering new insights!

Warm data labs provide a process to uncover these, and a space to allow for deeper systemic patterns within the collective ‘human information’ held in an organisation to emerge.

But the effect is much deeper than that — warm data labs also provide the conditions needed for people to adapt to a changing work environment. While many may fear digital transformation as a threat to jobs, this process may not only help them adjust through the transition, it may uncover the invaluable knowledge assets they bring to the workplace that are usually overlooked.

Where once organisations tried to align their people with their digital transformations, warm data may allow for a ‘digital evolution’ — so that the transformation evolves through mutual learnings between people and technologies.

Data Science

The quest to derive meaningful insights from big data has seen huge investments in developing data science capabilities and progressing the field. The ‘big data ecosystem’ is indeed a complex one, and for all the efficiencies that more data and smarter algorithms can provide — there remains a gap between data-driven insights and delivering better solutions.

Organisations recognise that big data is not the ‘silver bullet’ — and are using different approaches for delivering better big data outcomes, such as using ‘more nuanced small data’, focusing on building teams of generalists rather than specialists, and finding better ways to communicate data-insights amongst others.

Yet, some of the biggest difficulties faced by data scientists today are about how the data and models they use relate to the business or societal issues at hand — understanding and defining the problem (or sub-problems) being solved, recognising the limitations of the data models including data biases, and how data insights may inform decision-making are just the tip of the iceberg!

At the crux of the issue, what most companies are reaching for are better solutions for running their businesses or solving problems — and this involves both data science and people. But it goes deeper than that — it also involves how data science is influenced by the information that people hold.

The idea that information held within our individual experiences might allow for a deeper understanding into value-creation and problem solving is not new in data science — with most data scientists recognising the need for domain knowledge and multidisciplinary skills. Indeed, some of Google’s recent approaches to forecasting involve having ‘humans-in-the-loop’, thereby integrating data science and business planning.

What is new here is the ‘extra maneuvering space’ that warm data labs offer for uncovering new and deeper human insights, by allowing for the problem to ‘freely move’ across many contexts. Sense-making is enriched through a trans-contextual understanding of the interdependencies within the problems we otherwise spend so much time de-coupling...

In a previous post, we explored the possibility that working with ‘warm data’ could complement more traditional ‘data-driven’ approaches, and help uncover new insights into complex problems. And there are plenty of opportunities in data analytics to do so — one has only to encounter missing data, sparse data, data anomalies or seasonality data to see the immediate potential.

How warm data may influence the creation of data science models, and the judgements and assumptions we make around them will play a large role in their integrity and effectivess in organisations and society.

Warm data lab group processes also support communication through ‘mutual learning’ within the organisation. Data visualisations may be more effective when developed within these conditions — rather than at the end of a data science project.

Working with people on complex problems in a trans-contextual setting not only allows for deeper insights into the complexities within the system, it also creates the conditions required for shifts in perceptions and strengthens the bonds within an organisation’s ecosystem.

This article is based on “The subtle art of data gathering with Warm Data”.

Phoensight is an international consultancy dedicated to supporting the interrelationships between people, public policy and technology, and is accredited by the International Bateson Institute to conduct Warm Data Labs.

A Digital Evolution: Why Big Data needs Warm Data was originally published in Phoensight on Medium, where people are continuing the conversation by highlighting and responding to this story.