The Next Frontier Is The Self

Download and explore the pdf here

A decade of thinking about technology, identity and who we become

For years, we've asked what technology will do to our jobs, businesses, institutions and societies. I've been asking a different question: what is it doing to us?

Over the past decade, I've watched digital platforms turn our behaviour into profiles, algorithms use those profiles to predict our choices, and AI begin to recreate our voices, reconstruct our memories and even simulate versions of our future selves. These technologies don't just change the world around us. They increasingly influence how we think, how we see ourselves, what we remember and who we imagine we might become.

My work has followed this transformation, from our ability to imagine the future to the ways our identities are captured, represented and increasingly shaped by machines. Different technologies, but one enduring question: how much of who we become remains ours to decide?

Looking back across a decade of writing, research and public conversations, I can see how one question led to another. And how the technological frontier I've been following has gradually moved inward.

The next frontier is the self.

Before The Future of You: data and dreams

On 15 November 2016, I gave a talk called "Timing Is Everything" at TEDxWhitehallWomen, held at the Foreign & Commonwealth Office in London. It was about time, decision-making and our relationship with the future. The self was not yet the explicit subject of my work in the way it is today, but the shape of the problem was already there.

I opened with the proposition that "the future is one of the best decision-making tools we've got."

What worried me was our growing obsession with the now. The acceleration of technological culture seemed to be compressing our horizons. Everything was immediate, on demand and increasingly automatic. And under that kind of time pressure, I argued, people tend to do one of two things. We act on impulse, or we outsource the decision. More and more, I suggested, we would outsource it not to somebody else but to something else, like an algorithm.

That was 2016.

Even then, I was less interested in whether machines could make decisions more efficiently than in whether efficiency was the same thing as effectiveness. I drew on prospective psychology and the evidence that people decide better when they keep considering what might happen next. My argument was not that we needed to slow everything down. It was that we needed to see more time: to hold several horizons in view at once, and to bring the future consciously into the decisions we make now.

I ended with a distinction that has stayed with me ever since:

"The past can give us data, but only the future can give us our dreams."

And finally: "Not just to rely on big data for our big decisions, but to rely on our big dreams too."

From where my work stands now, that distinction feels more significant than I understood at the time. Data tells us what has happened. It finds patterns, produces evidence and can be used to infer what might happen next. Dreams belong to a different category altogether. They are about possibility, imagination and desire, about what has not happened yet.

So the issue was already one of agency. If we come to depend on machines trained on patterns from the past, what happens to our capacity to imagine what does not yet exist?

That was not yet Me:chine, or System & Self, or the Database Self. But the tension is recognisable. Data and dreams: what can be inferred from what has been, and what can still be imagined into being. The patterned past and the possible future. The system and the self.

The language would change. The question would deepen.

From prospection to identity

By the time I wrote The Future of You in 2021, the inquiry had moved explicitly into identity. The question was no longer only how people think about the future. It was what happens to the person doing the thinking once identity itself becomes technological.

One of the first changes I was trying to describe was distribution. Technology had begun to loosen the idea that identity sat neatly inside a single person and travelled with them as one coherent unit.

Social media made this immediately visible: a professional self here, a social self there, a romantic self somewhere else. Images, purchases, searches, credentials and relationships were scattered across networks. In the book's introduction, Distributing You, I wrote that digital technologies were making it possible to present different personas through different media and distribute our "self". The book was already asking what that multiplicity did to the idea of one coherent identity.

The self was becoming networked.

That did not automatically make it false. Multiplicity could also bring freedom. People could try out different aspects of themselves in different contexts. Avatars could become rehearsal spaces, and virtual environments could let people meet versions of themselves that physical life had not yet made room for.

By the end of the book, I was thinking about identity as something closer to polymorphic code. On page 190 of The Future of You, I wrote that we would build up a database of our own "characteristics, locations and relationships" that could be combined in different ways depending on context, and suggested thinking of identity as polymorphic code: changing form while keeping some continuity underneath.

That raised a question which has stayed with me ever since. If I can become different versions of myself across different contexts, what makes those selves mine? And if I become a different person across time, what connects the person I am now to the person I might become?

Self-continuity

That is why I have been interested for years in self-continuity.

As a futurist, I have naturally been drawn to our relationship with our future selves. But the idea is bigger than that. There is a past self, a present self and a future self, and some feeling of connection between them is what lets those different versions of a person belong to the same life.

I was already writing towards this in The Future of You. On page 87, discussing avatars and virtual personas, I wondered whether technology might let us meet different versions of ourselves at different stages of life, even to exist in some sense not only in the present but in the past and possibly the future. I imagined an identity capable of travelling through time.

The technology was primitive compared with what exists now. But the question was there: what if technology could put us into a relationship with another version of ourselves in time?

That matters because so much of life is a negotiation across time. The present self spends or saves, and the future self inherits the consequences. The present self learns or postpones, and the future self inherits the capability. The present self makes a promise, and the future self discovers whether it was kept.

This is why future self-continuity matters so much for pensions, wealth, health and any other area where decisions made now play out much later. If my future self feels like a stranger, it is easier to discount her interests. If she becomes vivid enough to matter to me, the calculation changes.

And technology can now make that possible self far more vivid.

Brands have always promised a future self

This part of the story is not entirely new. Brands have been in the future-self business for a very long time. As I said recently to the Swiss communications agency Farner in Zurich, brands have always been brilliant at promising us versions of the person we might become.

Wear a Rolex and become an admired self. Use L'Oréal and become a beautiful and worthy self. Drive a BMW and become a successful self. Fly Virgin Atlantic and become a more adventurous self.

The product is material, but the promise is transformational. It never only says have this. It says become this. We consume partly in the direction of who we hope to be, which is why the possible self matters so much to understanding both brands and futures. A brand can make a future identity desirable in the present.

But something changes when computational systems enter that relationship. A brand traditionally promised a future self. AI can begin to instantiate one. It can draw on evidence about me, construct a possible future version, give that version a face or a voice, and put my present self into some kind of relationship with it.

The imagined future self acquires an interface.

The Identity Century

By 2023, I had become convinced that identity was turning into one of the defining subjects of technological change. In August that year I called this the Identity Century, and my formulation was simple:

"Property defined the 20th Century. Identity will define the 21st Century."

The biology and psychology of the self, I argued, were being joined by a third dimension: the technology of the self. Technology was not simply giving us new tools to express identities we already had. It was changing the conditions under which identities are formed, expressed, authenticated, interpreted and governed.

The closer technology came to the person, the more technical questions turned into identity questions. Who am I to the system, and who can alter that? Which version of me counts, and what can be copied or generated? And who has the authority to speak or act in my name?

The Identity Century was never just about digital identity. It was about identity becoming technological. That proposition feels more relevant now, not less.

From distribution to reassembly

The distributed self led to something else. Once enough fragments of identity existed across digital networks, systems could start putting them back together. But they did not necessarily reassemble the person another human being would recognise. They reassembled the person a machine could understand.

That is what I came to call the Database Self.

I first presented it publicly on 3 April 2025, at the TBD Conference "Eigengrau" in London, in a paper called "Database Selves: Identity in a World Where All Data Is Identity Data". The Database Self is not my social media profile, or the version of myself I deliberately design. It is the operational version of me, assembled from accumulated evidence: something I bought, somewhere I travelled, someone I know, a search, a credential, an inferred characteristic, a probability. Different systems choose different fragments, weight them differently and assemble different versions of the same person. The bank meets one version, the insurer another, the employer another, the border another. An AI agent may assemble yet another, and nothing requires those versions to agree.

That is why I describe the Database Self as plural and programmable. In The Future of You: Five Years On, I described identity as increasingly editable, portable, searchable and programmable, and the Database Self as assembled through behaviour, transactions, relationships and inference.

The progression matters:

The network distributed the self.
The database reassembled it for machines.

Distribution created plurality. Reassembly created operational identity. The self no longer simply existed across networks. It became computationally actionable.

Database Brand

The same logic then reached beyond people to organisations. If systems can assemble a version of a person from accumulated evidence, why would they not do the same with a brand?

That became the Database Brand.

The Designed Brand belonged to an era in which an organisation could largely control the identity it projected through its logo, campaigns, tone of voice and brand guidelines. The Database Brand lives in a different environment. An AI system may build its understanding of a company from product pages, reviews, complaints, founder interviews, press coverage, sustainability claims, customer conversations, structured data and anything else it can retrieve and connect.

The result is not necessarily the brand as intended. It is the brand as inferred.

This is why Database Self and Database Brand belong together, as parts of a wider condition I call Database Identity. Both describe the same movement: from designed identity to assembled identity, from declaration to inference, and from controlled expression to distributed evidence read by machines.

The brand work brought the point home in another way too. Brands once promised possible selves to people. Now AI systems can assemble versions of both the brand making the promise and the person receiving it. Identity has become relational at computational scale.

From being seen to being read

All of this creates a distinction I now think is fundamental.

Human identity has traditionally been about being seen. We present ourselves, other people interpret us, and we respond. We explain ourselves, correct a misunderstanding, show up differently, perhaps change somebody's mind.

The Database Condition raises a different question: how am I read?

A machine-readable version of me may exist across classifications, scores, associations and predictions I never see. I can become highly legible to a system while its interpretation of me stays opaque to me. So the question is no longer only whether the data is accurate. It is what happens once the system's reading of me becomes consequential. Which version of me does the system meet, and what is that version allowed to do? Which opportunities are attached to it, and which quietly disappear? And what happens when an identity assembled from my past starts to shape my future?

This is where identity becomes a futures question.

When possible selves become computational

In 2024, I interviewed Pattie Maes and Pat Pataranutaporn from MIT about their Future You research. Their work makes a possible future self tangible and conversational. Instead of imagining some vague older version of yourself, you can talk to one a system has generated. That interested me immediately, because it connected identity back to self-continuity.

In 2026, I returned to the question in a Me:chine Dialogues conversation with MIT professor Sherry Turkle, whose 2026 book, Artificial Intimacy: Who We Become When We Talk to Machines, examines our changing relationships with conversational machines. What fascinated me was never just the technical accomplishment. It was the relationship: present self, future self, perhaps past self too, and the dialogue between them.

This is very different from conventional prediction. The message is not this is what will happen to you. It is this is somebody you might become. That distinction matters. A possible self should open the future, not close it.

But representations do not have to be literally true to affect us. Fiction affects us, advertising affects us, and so do memories and other people's expectations. A vivid enough computational version of a possible self can therefore influence the choices of the person it was generated from.

And that creates a feedback loop. A system observes me and assembles a Database Self from what it can know. It infers what I may do or become, and shows me a possible future. I respond, my response becomes new evidence, and the system updates.

At that point, the representation is no longer standing outside the person, describing them. It has entered the causal process.

A system's model of me can become one of the causes of the person I become.

That is one of the most important changes now under way.

When the map acts on the world it describes

We are used to thinking of models as representations: there is reality, and somewhere else there is a map of it. But computational systems act through their models. A credit model changes what somebody is offered. A recommendation model changes what they encounter. A risk model changes how they are treated. An AI adviser may shape a decision before the person even believes they have made one.

The map starts to act on the territory. And when the territory is a person, the implications are profound. If a system's reading of who I have been shapes what is available to me now, it may change which future becomes easier for me to inhabit next.

The issue is no longer simply whether machines can predict us accurately. It becomes:

What power does prediction acquire over possibility?

That question led me directly to System & Self.

System & Self

At some point, it was no longer enough to study the self on its own. The Database Self had revealed something larger. There was now a self assembled by systems and a self experienced from within.

The first could be made legible through attributes, records, patterns and probabilities. The second lived a life. It remembered, interpreted and imagined. It changed its mind, contradicted itself, refused and surprised. Neither could be properly understood without the other.

That relationship sits at the centre of System & Self. The system sets the conditions. Technologies, institutions, markets, rules and networks shape what appears possible, permissible or normal. But the self meets those conditions and responds. It adapts, complies, resists, improvises and refuses, and sometimes it authors something new. The future is made neither by systems alone nor by some fully autonomous individual standing outside them. It is made in the relationship.

Hence my core proposition:

Every future is written between the system and the self.

This is not a binary. System and self work more like figure and ground. Sometimes the system becomes overwhelmingly visible; sometimes the self comes forward. What matters is the relationship, the tension and feedback between them, and the question of agency within it.

Too Much Future

My more recent work on Too Much Future and Big Futurity grows from the same concern.

We are living through an extraordinary expansion of institutional futurity. Governments, consultancies and corporations all produce futures, and AI systems can now produce them too. At the same time, personal horizons can feel strangely small. The big future is AGI, climate transition and civilisational transformation. The personal future is whether I will still have a job, whether I can afford my home, and who I am becoming.

The collective future grows while the personal future contracts. We can become saturated with other people's futures and exhausted in the making of our own. Yet every vast systemic future still has to land somewhere: in a job, a home, a body, a decision. Every future eventually has to be lived by a self.

This is why System & Self is not just another framework for analysing change. It is a way of restoring the self to the future as a participant. Agency does not mean controlling everything that happens. It means keeping enough capacity to perceive the conditions around us, judge what matters and take part in what comes next.

In a sense, that takes me right back to the TEDx stage in 2016. The answer to acceleration is not necessarily more time. It may be seeing more time, and keeping possibility open.

From relation to dialogue

This is where Me:chine begins.

I first publicly introduced Me:chine on 2 October 2025 at the EDAY Digital Business Conference in Amsterdam, as the synthesis of the machinable and unmachinable self. Later that year I developed the argument in "The Sovereign Self", published in Five Critical Essays on AI, edited by Austin Williams. The name came first. Its full meaning only became clear once System & Self had made the relationship explicit, and showed that the encounter between system and self is becoming active and continuous.

The system reads me, and I respond to how I am read. It anticipates, and I adapt. It recommends, and I accept, ignore or resist. Then it learns from my response, and the relationship becomes recursive.

In Five Years On, I describe Me:chine as the place where systems and selves continually meet, interpret, negotiate and alter one another. There is a simple way to describe the whole progression:

The network distributed us.
The database reassembled us.
AI puts the assembled self back into conversation with the living one.

That conversation is Me:chine.

The colon

The punctuation matters.

Me : chine.

The colon separates, and it connects. Me on one side, machine on the other. Neither collapsed into the other, and neither held completely apart. It marks a relationship, not a destination.

That is why Me:chine is not a man-versus-machine proposition. I am not interested in which side wins. Nor am I interested in a story where the human slowly dissolves into the machine until the distinction stops meaning anything. The interesting territory is between.

The colon can hold several relationships at once: present self and future self, the self as experienced and the self as represented, what I am and what I might become, prediction and possibility, the machinable and the unmachinable.

The colon is not empty space. It is where the negotiation happens, where each side acts on the other, and where new forms of selfhood are beginning to emerge.

Why the Me:chine Dialogues are dialogues

This is also why I call the interview series the Me:chine Dialogues.

The conversation between me and the person sitting opposite is only its visible form. The deeper dialogue is the one inside Me:chine itself: between the machinable and the unmachinable, between what systems can capture, classify, reproduce or predict and whatever escapes that account.

Guests come to that dialogue from very different places, from technology and philosophy to psychology, art, memory and intimacy. But the underlying question stays the same: what happens when technology begins to take part in the making of the self? Not merely to observe it, serve it or represent it, but to participate in making it.

That is why Me:chine is best understood not as a conclusion but as a continuing condition of inquiry. The relationship keeps changing, so the dialogue has to continue.

Not Man versus Machine

This is the distinction I want to hold on to. The future is not best understood as man versus machine. It is where the machinable and unmachinable selves meet.

That does not mean drawing a permanent line around the things machines will never do. The line will move. Machine capability will expand, and more of the person will become legible, reproducible, predictable and computationally actionable.

But that only makes the question of the self more important. A copy is not automatically the whole thing. A representation is not the life it represents. A prediction is not a destiny. And machine-readability is not the same as being human.

That is why the language I have used for Me:chine comes back to a meeting of selves:

Not a war between man and machine,
but a meeting of selves.
The part that can be copied,
and the part that never can.

To live as Me:chine
is to stay awake inside the system,
to hold what is machinable,
without surrendering what is not.

That, for me, is the frontier.

The Next Frontier is The Self

In August 2023, I argued that this would be the Identity Century. I think we are only now beginning to see what that really means.

People and industries that once thought mainly in terms of technology are now moving towards questions of identity, agency, continuity, intimacy and selfhood. Not because they have changed subject, but because technology has reached the subject. It has moved from the device to the network, from the network to the profile, from the profile to the database and from the database to the model. And now the model is moving back into the life of the person being modelled.

Possible futures are becoming possible selves. Brands have always promised those selves. Systems can now infer and simulate them, and AI can put them back into conversation with us. The more machine systems shape what we encounter, what we are recommended, what we delegate and what seems possible, the harder it becomes to pretend that technology is something outside the person.

The technological frontier is moving inward. So the next frontier is not simply artificial intelligence.

The next frontier is the self.

The self in both its machinable and unmachinable dimensions. The self distributed and reassembled, known and acted on by systems, and still capable of responding, resisting, imagining and becoming. The self in dialogue with the machine.

Which is why, looking back over the past decade, my work now looks less like a collection of separate projects and more like one inquiry growing steadily more precise.

Prospection became identity.
Identity became distributed.
Distribution made reassembly possible.
Reassembly made the relationship between system and self unavoidable.
And that relationship has become a dialogue.

The Future of You asked what technology was doing to identity. The Database Self asked what happens when systems can reconstruct us for themselves, and The Database Brand showed the same logic reshaping institutional identity too. System & Self asked what happens between those systems and the selves living within them. Me:chine asks the question that follows:

What happens when machines begin to participate in the making of the self?

That is the question I expect more and more of the technological future to converge on. Because the Identity Century is not simply about proving who we are.

It is about deciding who we become.

A timeline of the work

15 November 2016: "Timing Is Everything", TEDxWhitehallWomen

At TEDxWhitehallWomen, held at the Foreign & Commonwealth Office in London, and speaking as Chief Strategy & Innovation Officer at The Future Laboratory, I argued that "the future is one of the best decision-making tools we've got." I explored how an accelerated, present-focused culture encourages impulsive decisions and the outsourcing of decisions, more and more often to algorithms. Drawing on prospective psychology, I argued not for taking more time but for seeing more time, and closed with the distinction: "The past can give us data, but only the future can give us our dreams."

The central concern was already agency: how do people keep the ability to imagine futures and act towards them, rather than letting decisions be generated entirely from patterns of the past?

2021: The Future of You

The inquiry became explicitly about technological identity. The book described the distributed self, networked and multiple identities, machine-readability, biometrics, avatars and synthetic selves. It speculated that technology could let us meet versions of ourselves across time, and ended by describing identity as polymorphic code, assembled from a database of characteristics, locations and relationships.

August 2023: The Identity Century

I argued that "Property defined the 20th Century. Identity will define the 21st Century." The biology and psychology of the self were being joined by the technology of the self, and identity was becoming one of the defining terrains of technological change.

2024: Future You and self-continuity

My conversation with Pattie Maes and Pat Pataranutaporn about MIT's Future You work took the inquiry into computational possible selves. A future self no longer had to remain purely imagined. It could become tangible and conversational, and self-continuity became something technology could mediate.

3 April 2025: Database Self

At the TBD Conference "Eigengrau" in London, I presented Database Selves: Identity in a World Where All Data Is Identity Data. The distributed self became something systems actively reassemble: an operational version of a person, assembled from accumulated evidence, machine-readable and capable of being acted on. The question shifted from how do I present myself? to which version of me is the system reading, and what can that version cause to happen?

2 October 2025: Me:chine

At the EDAY Digital Business Conference in Amsterdam, I publicly introduced Me:chine and the distinction between the machinable and unmachinable self. Later that year I developed the argument in "The Sovereign Self", published in Five Critical Essays on AI. It moved beyond man versus machine to a different question: what kind of self emerges when the machinable and unmachinable dimensions of identity coexist in AI-mediated life?

Five Critical Essays on AI, ed. Austin Williams, TRG Publishing, 2025. ISBN 978-1-903094-09-9.

April 2026: System & Self

The inquiry became explicitly relational. If systems assemble operational versions of us and shape the conditions in which we act, where do agency and authorship remain? System & Self provides both the method and the proposition: every future is written between the system and the self.

June 2026: Database Brand

With The Database Brand: How AI Is Changing What a Brand Is, the same logic extended from people to organisations. The Database Brand is the version of a brand that machines assemble from accumulated evidence, marking a shift from designed identity to inferred identity.

2026: Me:chine becomes a dialogue

The Future of You: Five Years On and the Me:chine Dialogues took the idea further. The relationship between system and self is no longer static. Systems and selves continually read, respond to and alter one another, and Me:chine has become an ongoing programme of inquiry into what happens when machines take part in the making of the self.

Now: The Next Frontier is The Self

The technological story has moved from devices to networks, from networks to profiles, from profiles to databases, from databases to models, and from models back into the lives of the people being modelled. The next technological frontier is not simply the machine.

The next frontier is the self.