Human Resilience in the AI Age

There is a version of the AI future that does not arrive with a bang. No conscious machine announcing itself. No single moment when the meteor Hits. <LYRICS>Don't come away with me.</LYRICS> Instead, it arrives quietly. A recommendation here, a risk score there. A route selected, a sentence completed. A candidate ranked, a child taught. A neighbour informed by a different reality from your own. And slowly, almost politely, artificial intelligence becomes the thing beneath things. An operating system for decisions. Relationships, institutions, and everyday life. My guest today is Lee Rainie, Director of the Imagining the Digital Future Centre at Elon University. For more than two decades, Lee and Jana Anderson have been asking experts to look around corners and over horizons, producing more than 50 reports on the future of digital life. They began in the early days of broadband, then mobile connectivity, then social media, and now they are tracking what Lee calls the fourth great turning of digital life: artificial intelligence. Their latest report, Building a Human Resilience Infrastructure for the AI Age, asks what may be the most important question of all—not simply what will AI do, but what will humans need in order to survive, adapt, and remain capable as AI becomes woven into the systems around us. In this conversation, Lee says something crucial. The challenge may not be one big artificial general intelligence moment. It may be the quiet takeover of systems and decisions. AI insinuating itself into the structures that already mediate our lives, work, education, finance, healthcare, government, media, and more. And that is where this conversation meets the heart of me sheen. Because the future is not only about smarter machines, it's about altered humans. Lee and I talk about resilience, but not as grit, not as the old idea of simply bouncing back. This report argues that resilience now has to become infrastructure, institutional, civic, educational, psychological, social, and ethical. We talk about existential literacy, about friction as a human good, and about why the future of agency may depend on our willingness to pause. This is not a conversation about rejecting AI. It is a conversation about what must not be handed over. In my language, what is unmachineable. Because if AI does become the environment, ambiently so, then the question is no longer whether we will use it. The question is whether we can remain human inside it.

Welcome, Lee. It's so fantastic to have you here. Thank you. I know it's early where you are, although I know you're always an early bird, but you're early with everything. You are the early warning system on AI.

That's our aspiration. We talk to experts like you to try to figure out where we're going, and that's our goal.

Yes. Well, I followed your reports and your research work, both you and of course Jana. Janna Anderson, who I met, I think, about 10 or 11 years ago. And you're both fantastic. I follow your work and try and share it as widely as possible alongside lots of other contributors to your reports. But I wonder for people who aren't that au fait with yours and Janna's work here, whether you could just give a little bit of background, how Pew and Elon have come together and the sorts of subjects and the way in which you research them has continued to evolve.

I was hired at a think tank in Washington, the Pew Research Center, at the turn of the millennium. And at the same time, my daughter was a student at Elon University in North Carolina. So I got to know the faculty there as I visited campus. They were interested in the work I was doing. I had also just come from a magazine gig where I was the managing editor of a Newsweekly magazine in the United States that was famous for doing college rankings. And so I was treated as royalty when I came on campus in the hopes that the rankings of the school would improve. And it did, but it had nothing to do with me and my daughter. It had to do with the school doing really well. But it put me in touch quite early on with the faculty, especially in the communications department, the people who study media and strategic communications and information ecosystems. And soon enough, Jana Anderson and I, a professor at Elon, discovered each other. And she's a workhorse. She's just a great researcher, ambitious, and very indefatigable. She just wants to do lots of things. So we figured out we could do work together really quickly. So really, for 23 years, we've been putting out these Future of Digital Life reports. And in the early days, we were asking about what's the future of mobile connectivity and what's social media gonna do to us? And is the web gonna be overtaken by apps? And what what's going to happen with extremist groups. So we have this long-running series of reports. The one we're going to be talking about today was the 52nd report that we had done together. And we eventually got pretty good at sort of seeing where conversations were going. We're seeing where predictions were going about the internet. And we wanted to talk to experts like you about what they saw around the corners or over the horizon. So that sort of mindset informed a lot of our work. And I thought I was going to retire from my career when I left the Pew Research Center, but then the president of Elon called me up and said, do you want to keep doing this work with Jana? And I said, sure. So that's why I'm now housed in the Imagining the Digital Future Center at Elon. And since mid-2023, and that's when we really became sort of enchanted with you and your work and begged you basically to answer our questions. And so we've been doing a lot about artificial intelligence, the sort of the fourth revolution in digital life we've covered. The first was sort of early days broadband. You know, everybody thought that was radical at the turn of the millennium. Then we studied mobile connectivity, then we studied social media, and now this is the fourth turning of the wheel. So we're kind of lucky social scientists just for doing that. And it became really obvious that the big direction we should go is what's the fate of humanity as artificial intelligence, particularly the Big Bang moment was November 30th, 2022, when ChatGPT came into the world and everybody became aware that these systems were growing. So that's how we produce these reports about the future of human capacity, as well as now this newest report about resilience strategies for humans to make sure they survive the onslaught of AI.

One of the things I've always loved about your reports, even before I was a contributor, which it is, by the way, an honor to be, is the blending of the, you know, the real executional, practical stuff that people talk about and that you include in the conclusions and the findings, and also the civilizational context. You know, it's so hard to put that macro and micro together, or the conceptual and the practical together, but every single time you and your reports always do it. And the one The one that you just alluded to that came out in April 2026 is Building a Human Resilience Infrastructure for the AI Age. It seems very different to some of the previous reports. So for example, Being Human, got a lot of stats in it, qual, quant, drilling down into work and whether people think, or the experts that you've interviewed or who've been surveyed, think human work is going to be replaced by AI. Are they positive, negative, optimistic, pessimistic? This is much more conceptual and you've arrived at this concept of resilience. I wonder if you could explain a little bit about that and why you thought that that was, you know, the big sort of finding or outcome.

We beginning a couple of months before we field the survey to be thinking about the main idea that we're going to pursue and we are not driven by an agenda. We're driven by where our sort of gut sense and our reading of the landscape shows us that interesting conversations are emerging. And we don't want to do me too work. We want to push the envelope. We want to be at the edge and we want to think about things that people haven't yet really thought about in a grand context. So frankly, Tracey Follows, it was your answer in 2025 that pushed us to pursue the resilience idea in 2026. I will remind you that your answer when we asked about human capacities, we were asking about What is AI going to do to creativity, to curiosity, to social intelligence, to critical thinking, all the sort of basic stuff that had been talked about for a long time about how are humans going to survive the AI onslaught? And it was your ideas about the merger of machine life and human life and how to sort out the distinctions between machinable, as you call it, and un-machinable aspects of life And it felt like it was a time for us to stop documenting people's worries and concerns and begin to think about strategies or capacities humans have to think about doing well as AI encroaches on so many more aspects of human life. And it was so interesting to just do the literature review on this. I had a vague sense about what might be said, but it was just so astounding to me for reasons that I probably should have anticipated. When you start asking about resilience, you're talking about a lot of painful things. You eventually get into the literature on PTSD. How do people survive deep trauma in their lives? You get into the literature about grief and who doesn't do badly as they're in the midst of a grieving process and those who just get completely demoralized and lose faith and hope and capacity to rebound. And so it was an interesting exploration in the first place. And one of the things we tried to do in the survey was explore different dimensions of resilience. And that's where the magic came together in this report. And frankly, there were just really fascinating and beautiful surprises in these data. And we can go through those, but it was just a different kind of learning experience. The other thing that it felt like an okay moment for a center like ours to really come up with something. It wasn't a manifesto, but it was close enough to sort of saying, here's how to think about this and work on it, not just sort of what are the diagnostics of what's going on. And it was a very different kind of report. I'm glad you picked up on that because it was given in a different voice than our previous reports. It was treated as a mini manifesto. And it was meant to launch new conversations about, okay, now we know this is going to happen to us. We know that our capacities are going to be challenged. So what do we do about it? And writing in that voice was a really interesting thing for us to try to do.

How do you think the notion or concept of resilience has been received by those who weren't contributors, who are outside of it, maybe in PR or the media or other academics who will obviously be following the findings of your reports? Tell me a little bit about whether it's been different. Maybe it's been the same. I don't know.

It has been different. And in part because of the things I've already described, it felt like a— your invitation a couple of years ago was, okay, let's think about solutions. And we hadn't explicitly gone after that in the way we did directly in this report. So I think the reaction to it broke into some really interesting distinct camps. There were some people who were just happy that we had something to say about solutions and something to say about a concept that moved the debate to a different plane. It wasn't just hand-wringing about the things that were going to happen, but it was thinking about strategies. Other people were really enchanted with or interested in the idea that we took as a surprise and wrote as basically the sort of headline overarching framework and concept that this report, which is that traditional notions of resilience and things that scholars and other experts had been looking at for generations, was not sufficient to be thinking about the future with AI in it. Their argument was, yes, of course we'll have to be adaptive. Of course, the capacities for grit, for absorption of challenges and moving on from them The capacity just to persevere is always going to be important when you're thinking about resilience. But the real shocker to us was how many different voices, including yours, referred to institutional needs that are arising in this era. The basic argument was AI systems are now mediating so much of life. They're regulating our rights, they're regulating our risks, they're mediating the information that we get and things. That it will take a systemic response to those challenges rather than an individualistic response, because you just need the force of institutions. You need the collective expertise of institutions. You need the enforcement mechanisms of institutions to push back against some of the most troubling challenges that AI has raised. It was fun to write that up because it was just not the thing we were expecting when we first went into the field with this survey.

What's the age profile of most of the contributors?

They're sort of middle-aged experts from a variety of fields. There are people like you who are just sort of in applied critics and pundits and operational people who are trying to help people navigate this world. If the database we've developed, it's not representative, just we should say as we're talking about this, this is not a scientific sample of expertise. It's sort of people we respect and we've watched and people who have interesting things to say about technology, some of whom fit your profile, just experts in the field doing the work. Some of them are scholars, a lot of scholars who are more from a critiquing and punditry perspective. Some of them are the pioneers of the internet itself. I mean, Vint Cerf, Doug Engelbart, other people have been respondents to our surveys for forever, and we're so grateful to them. Some of them are building the technologies. It was really interesting to see that we had some folks here, but both sort of inside some of the AI companies and then people who had left AI companies partly because they're concerned about the future. So it's a, it's a very diverse sample in its way, more male than female. We allow people to be anonymous in the surveys because sometimes we know they're inside companies where they aren't following the company line as they write their answers. So they don't want to get in trouble and they don't want to have their words misinterpreted as the official position of the company they work for. And it's a pretty North American sample, a good response from Western Europe in particular. But we get a smattering of Asian and respondents from Oceania. We've had a much more robust response rate in Eastern Europe this time than we have before. So it's a mixed bag, but it's, you know, people like us more or less.

That's great background actually to that. I wondered about age because of the importance of institutions, as you say, that came out very clearly in the report. You know, this almost assumption that somehow, some way institutions can crack this if we just help them enough. And I wonder if like a younger generation are like a bit over institutions or they want their own institutions, like we've seen the growth of DAOs, decentralized sort of versions of governance. And so it's one of the interesting things I think, is there an assumption that, I think in the report it comes up quite a few times, you know, global institutions for example. And of course at the moment we're seeing the proliferation of the world into multipolar sort of spheres of influence. So the institutional question's a good one, isn't it?

Yes, it is. And you're right to pick up that younger folks just as a rule are more skeptical more wary, more likely to say that institutions, just what are they? They're so old school and many of them are so hidebound. It's a synonym with bureaucracy and red tape rather than the glue that holds civilizations together. But it was interesting here to see that there were eclectic answers on this. There were some people who talked about distributed institutions, There were some people who talked about applying sort of the better practices from the modern era in collecting information and doing crowdsourcing of information that thought that institutions might be able to recover partly from that. But it's a tough sell, particularly in America, where the decline of trust in institutions has just been spectacular over the past two generations. So we're arguing upstream, or the experts that were giving voice to in this survey are arguing upstream against institution building as a civilization strategy or societal strategy because it's just, nobody is feeling very good about institutions these days serving people's purposes.

Because in the report, one of the findings is, and correct me if I'm wrong, is it 5 layers for institutional-first infrastructure? The second one I think that's in there is really interesting. It's that civic layer, which feels quite millennial in a sense, community-driven, participatory, thinking about your local area, shared values, collective deliberation. That feels to me like it was really bubbling up through the findings.

Yeah, good read. And to me, that's sort of the most exciting place of potential in these arguments because people don't necessarily think that civil society is in the kind of trouble that sort of formal institutional public governance structures are in trouble. And there are ways in which almost by force because things are so bad, particularly in America with institutions, that building stuff from the ground up just makes a whole lot of sense. I mean, we picked it up in other ways in previous surveys and the work I was doing at the think tank, the Pew Research Center. The only way Americans think trust is going to be rebuilt is neighborhood by neighborhood, shoulder to shoulder, person to person solving problems because You have toxic partisanship now is the dominant broad societal story. Nobody thinks that there's a way to break through that until people start seeing their neighbors as allies again. And of course, one of the big challenges of that, that's implicit— no, it's actually explicit in a number of ways in this report— is the epistemic fragmentation that we're going through. People having different realities that they're responding to, people having trouble navigating to the truth. People not having a sort of shared sense of what reality is, is one of the hindrances across the board on this, but particularly at the local level where you and your neighbor might live in completely different information zones. And in your estimation of this, the machinable parts of them are just so at variance, they might as well be living in two different galaxies. So there's a way that at least if you've got a local problem to solve, if the sewers need to be fixed, if there's a disaster in the local area where people literally have to help each other to survive, well, that's the sense that people are going to be forced to come together to solve local problems because they can't avoid them. And that might be the formula for better institutional trust to emerge.

I'm sure if politicians just let us all get on with it, it would be absolutely fine. That's just my, my view. You also have the cognitive layer. This is another theme that came up, I thought. I know everybody's jumping on the word literacy at the moment and attaching it to everything, but a kind of cognitive or AI literacy, but going further than just, you know, cognition or metacognition, actually thinking about the ethics, sociology, philosophy, some of the things that perhaps in decades and eras gone past, you would have called it human flourishing. How does human flourishing and a cognitive layer sort of sit together?

There were two things that were, again, sort of surprises to me coming out of these data that were so interesting to write up. The first is the almost across-the-board encouragement of these experts to build basically friction into processes that deliberate pauses, literally, that can be orchestrated by technology, but in some cases just by personal practices. That friction is our friend, these folks are arguing, and it is the thing that AI is pushing against, is meant to solve, where it's supposed to get us more efficiently to answers or to moving on with our lives. But there's so much argument in different contexts in this expert commentary that pausing, living with the messiness of humans dealing with each other, In the context of metacognition, just remember where your blind spots are. Remember where you might default to an easy answer rather than wrestling with the harder answer. Obviously, the call to check your sources very well. And so there's this huge argument for what the Greeks taught us, which is a concept of phronesis, which is practical applied wisdom. You know, you learn stuff in life and then you live it because you've learned from your mistakes or you've advanced your knowledge and things like that.. And so that was one of the calls here. The other one was the strong encouragement of these experts to expand the notion of AI literacy. I mean, you know, you got to know how to use the tools. I mean, that's job one, just because you need to know what you're talking about and need to know how to frame your arguments or to frame how you're going to live your life with these tools. But the sense that what they called existential literacy, is the grander call to us, that an array of institutions, an array of personal practices, an array of ways that we are civically going to be supporting each other is to teaching ourselves about the grandest parts of human existence, the stuff that you call unmachineable. You know, it's— it is metacognition. It is critical thinking. It is collective decision-making. It's kicking back against the systems that are sort of forcing you to move in one direction or another or trying to make decisions for you and things like that. There's a sense that a quite whole framing of how to live your life. I mean, basically back to what those Greeks were talking about and what actually one of the striking things about this work is that the people who have thought longest about resilience are people who are religious leaders in spiritual communities. Literally defining what communities are and how people ought to consider living their lives.

Yeah, that is really interesting. So let's get into it. Remind everybody, it's a big report. How many pages have we got?

300-something? It's 375 pages. We sent out an invitation to a couple of thousand experts and 386 responded to them. Again, a nice diverse sample, not the scientific sample, study. So we're not trying to hold out our quantitative data as, you know, absolutely dispositive. From all over the world. And about 160 of you folks wrote really incredibly powerful, thoughtful essays. And so sorting through those essays and finding the themes and patterns in them was the biggest job of writing the report on this. And frankly, AI made it better. I mean, it just— there was a way in which we could be thinking about the arrangement of the themes and the orchestration of the report and the order of the report. That AI was a big help in doing that sort of grand thematic first cut at the data. We turned it into our own stuff, but it, unlike previous years where we didn't have AI tools like this and didn't want to use AI tools like this, we now discovered that at least at the layer of organization and structure, AI was a really helpful part. And we asked you experts in our survey, you will remember, When we got your answers, the next question after that was, how much did AI help you write your answers? And a lot, about 3/4 said not at all, but a quarter of experts, folks who know this stuff, said it was a help and I used it and I'm not going to shy away from it.

Yeah. Yeah, I remember answering that question because I was hovering over thinking, I wonder how many people will be honest about this. I'll be completely honest. I mean, they're my ideas, but I find them so much better crafted with the help of AI that it can summarize, pinpoint, compare, especially when you are in it, when you've been in an idea or a concept and you have your own blind spots, don't you? And actually you want to say to somebody, can you objectively just sort this like a bit of a sorting hat? What's the priority point I'm making? And I'm sure the same for you with the analysis.

Well, and the other thing that I'm picking up, it's just sort of sitting in the academy as I do now at Elon in North Carolina, is that the most creative teachers of this stuff are encouraging exactly what you're talking about. You know, red team my answers, or show me where my blind spots are, or show me where my argument isn't convincing, and things like that. And they're finding that their students are becoming better thinkers and writers because of it. And a lot of professors now are sort of saying, as a matter of extra credit in the course, if you, student, teach me something about AI, I'm going to give you an extra credit. So it's a co-learning environment where that shows up in our report too. Yeah, a bit reverse mentoring.

We love that. So I mean, the chapters, I'm not going to obviously list them all here, but just to give people a flavor, they go from everything from prioritizing autonomy. I mean, agency, we'll talk about agency because it's the first chapter and it's close to my heart, but I think it comes out in the report just how important it is cultivating human agency. But obviously you talk about institutions, The Ultimate Team-Up, which we've just been talking about, humans and AI working together. Existential Literacy, which you also mentioned. Workquake, we'll come back to that in a minute. Navigating Labour Shifts and the Pursuit of Meaning. The Great Divide about the inequalities. Heart and Soul, you know, the thing about the human connection. Overcoming Complacency and the Lure of Convenience. I mean, that came through massively and everybody feels it in their everyday, don't they? And of course, misinformation, illusion, the epistemic vigilance, and lots and lots of broader insights closing off with what is the path to human flourishing. Can we go back to work quickly because you've done so much work on this through the previous reports and research as well. What did you find that was really surprising, or maybe nothing was, but really surprising about this navigating the labor shifts and the pursuit of meaning?

We've had a really interesting evolution on thinking about this. Our first expert canvassing on the subject was 2013. So it was right after the first wave of reports from Oxford and some other places about the potential for massive job loss. So we asked the experts, folks, are more jobs going to be created than lost or lost than created? And we had a 50/50 split. Half the experts said history is a good teacher here, that we've survived disruptions. It's been messy as the disruption occurs and it's caused a tremendous amount of human and community pain. But eventually we come out the other end with wealthier societies, more jobs being created, new jobs coming into being that were never possible before. And that should happen here. Those who pushed back against that answer basically said, this is a different moment from the past. We've got, you know, the nano-informatics revolution. We've got genetic revolution. We've got the cognitive revolution itself. We're studying the brain. And so at the same time, it's the term of art in the academic community is poly-crisis. There are so many things unfolding so fast on so many big fronts that, you know, humans can barely keep up with the biggest of them, the job-quake kind of things. But there are We just won't survive this onslaught and it will, and the disruptions that are going to be caused are going to be so disorienting that we basically won't have our wits about us and come out with solutions on the other end. So we move along, we move along. And even in that early survey though, people were saying the saving grace for humanity, whether you thought the job quake was going to break bad or break good, the saving grace is what was called soft skills. Remember that phrasing? It was going to be the community-based things, the critical decision-making in uncertain environments, critical thinking, social and emotional intelligence, and things like that. Well, soon enough, when AI began to show the capacity to look like it was doing things that were empathetic, that were exposing curiosity or somehow, you know, furthering creativity, this is now on the table. That's why we did the survey we did in 2025 about human capacity and followed up with this one. So what was new in this report was the sense that people's jobs are now so deeply bound up with their identities, their sense of purpose, the way they make meaning out of their lives. That's the thing to be thinking about. There's still expert commentary one way or another about whether the job loss is going to be greater or less than what we currently have. But there's absolutely no doubt about you experts, I would say, that the way in which people think about how to build their lives and how to make meaning out of their lives, and particularly whether their job gets lost or not, it's certainly going to be disrupted. And figuring out how much work they want to do to make themselves relevant for pay and how much they want to move on to other things, that, that's the play here. That's the thing that's most interesting to be thinking about as a future concern to address.

Exactly. And curiosity came out quite high, didn't it, in, as you mentioned, in the previous Being Human report. And I just wonder, where are we going with curiosity and AI? Obviously, AI is doing some amazing pattern recognition and it's delving into all of the training data that we are so lovingly providing it with. But it can't think about or invent things that it has absolutely no training data on. Where do you think we are, either from your own point of view or through what came out through the report on curiosity?

The new thing that we've done in these future reports since our 3 years at Elon is to add, where relevant, public opinion comparisons to the, you experts and your data. If you might remember, in 2025, the experts were pretty sure that the impact of AI and creativity would be to expand it. I'm just personalizing my sense of what the answer is because they had different reasons for thinking that. But my sense is, and I bet you this is the case with you too, when you have an itch on any subject, in any context, this can take you down the deepest rabbit hole that you want to go down. So it's not necessarily thinking up new stuff. I mean, that— I mean, which is a classic definition of creativity and a foundational one. But I'm thinking more like I'm sitting here as I'm writing the questions for this survey that we field at DU on resilience. And I wonder what the grief community has taught us about this. I didn't know, but I found out relatively quickly and could then go to some incredible literature that helped us frame some of the questions on this. This. I wonder what the difference is between individualistic resilience and community-based resilience. Obviously, communities have suffered big losses, disasters, you know, in wartime, you know, just the kind of horror that they're going through. What's the literature show on that? And again, AI is not a bad starting place as long as you don't get sort of tricked by it or as long as you don't think it's the be-all and end-all answer to things. So I was thinking about creativity in that way as I was trying to interpret these answers. And so I'm interested in kind of how you think about it. But the one other thing that was striking as we compared expert answers to general public opinion, just as average citizens, representative sample of American adults, American adults were deeply pessimistic about the impact of AI and creativity. So they're different. From you experts who had pretty good hopes for it, and across the board, all of these traits that we asked about in general public opinion, creativity was one of the things that average everyday folks thought, that's going to get hammered too. There's just no way AI isn't going to overtake basically all the essential human traits that we were talking about.

Yeah, that's upsetting, isn't it? That people think that and that it could also do that. I mean, when I've researched things in the past or I've been working on projects, not everybody wants to be creative or they need a little bit of help at the beginning because creativity is quite intimidating. So people need a starting point. And I wonder if actually, you know, the whole role of AI is actually maybe to be a little bit of a starting point for that, to give some framework so that people can be creative. I mean, I've always thought that AI, certainly in the short term, is not going to make us more productive because I haven't seen over these longitudinal studies in the workplace that anything of all of these tools, the productivity tools we've got, has made us more productive, particularly. You may disagree, but I've always thought it would make us more creative, just especially for people who are working on their own or they have an idea and they don't have anybody in the home or in the office to discuss it with. Yes, you can't replace, you know, the water cooler moment where something sparks a completely innovative idea that, you know, it's just very serendipitous. But imagine all of the things that go through one's brain in the day, a way of recording or linking them together, like having the, "Oh, I just thought of that in the shower" moment, but through having a chat with ChatGPT or Claude or whatever, surely is a good thing. And I really think at some level, I know there's a lot of talk about it really bringing in a lot of mediocrity, but I think for a lot of people, it's gonna upskill us in creativity.

Yes, and there are, in the two surveys that we're talking about of experts, I mean, one of the sort of pieces I would say of conventional traditional wisdom that now comes out of your world is that if done right, AI can be an incredible time saver. And so then the question becomes, what do you do with that time? And you can be an idiot about it and not creative at all and just laze your day away or whatever. But if you, if it sparks for you and it takes you down different pathways, all for the good. And so there's a There's a sort of call in these data for thinking about where AI creates abundance and the possibilities that abundance gives us that it didn't give us before. I mean, professional expertise, for instance, is now being commodified in some respects. And where does it give us sort of new capacity to do new things? Where are opportunities being opened up that only required, you know, high levels of education, high levels of professional training and things like that. Where can now everyday citizens do that and bring their own sensibilities to those processes? There's going to be a flowering in directions of where abundance frees resources and where people feel the capacity to move on to other things because they've got now some other system humming away in the background taking care of the things that they used to spend their time doing. Exactly.

Now, on the flip side of that, of course, we had the film Idiocracy quoted in the report, and I'm sure it was more than once, and lots of people pointed to the fact that, oh, it's going to make people stupid, or artificial unintelligence, or human stupidity, etc. Can you talk to that theme as well and what the findings were?

Well, the new thing there, I mean, obviously in the quest of human agency, there's a sense that deferring to AI has been a long-running conversation. We've picked up on it. We did previous reports on it. What the new piece of it that made sense once I read it was that humans, first of all, we're efficiency machines. We want shortcuts. And there are plenty of ways that the average human, even the discerning and caring human, you know, needs to practice heuristics and to know where to, you know, not waste their time doing things. And one of the tricks, I would say, in quotes maybe, that AI plays on us is to give definitive answers. Answers, or give answers that look definitive, look sufficient. If you're a satisficer, you know, that's good enough if you've got to hand in a report to your boss. And that the definitiveness or the seeming definitiveness of AI answers is going to be one of the real traps of it, that we will defer to it in ways that are just totally inappropriate. And so that was one of the elements of what you're talking about, which is the wall L.E. world where, you know, people are sipping their umbrella drinks at the poolside and just letting AI do everything for them. And again, sort of that goes back to this idea of friction finding or practices bringing friction to your work to make sure that you're in the loop and that your sensibilities and your perspective and your capacity to think through things is enhanced rather than handed off.

Yeah, 'cause a lot of people were talking about just how invisible it's going to become and that that makes it easier to sort of hand off things to it because I think you were talking of it in terms of being frictionless or friction-free, but that invisibility, I think that invisibility worries people.

Yes, very much so. Again, sort of that's the element of the important finding at the very beginning of the report, which is essentially, in many of these experts' estimations, the arrival of artificial general intelligence or artificial superintelligence, a big bang moment, the meteor is going to hit. And then, you know, the day after that, everything has changed. There was a much more subtle and interesting, and my sense of it was a reasonable argument that it's the quiet takeover of systems and decisions in the insinuation of AI into systems that already exist. That evaluate and mediate and make recommendations and make predictions about us. So there's that silent operating system that these experts were talking about when they were talking about rebuilding and newly building institutions that will have our backs. I mean, it was in some ways you can't keep track of all of this. So you need systems of accountability and insistence on transparency and enforcement mechanisms because it's just too complicated for individuals or even groups of individuals to pay attention to every element of this that is bringing itself into your life. So that was the institutional argument. And I'm glad you brought that up because it was, you know, these are scholars and folks with insights even in the tech world where it's not going to be when the meteor hits, that's the big moment. It's going to be the quiet takeover and the quiet ways in which it insinuates itself into our lives.

Yes, because towards the end of the report, You talk a lot about there are findings around physical AI and AI becoming, you know, part of an ambient system. This is what I mean by the invisibility. Obviously we're in a sort of halfway house. We're nowhere near that yet, but people can already feel it happening. And so they're talking about AI coming into smart systems, as you say, embedded into them where you'll have, you know, the sensors and the facial recognition and the biometrics all talking to each other to act maybe on behalf of citizens, maybe not, in real time. And I was looking, I mean, I've had for several years in my presentations that humanoid robot that they have on the streets of Bangkok, for example, you know, which is, it's all very well looking as sort of Optimus and all the things it could do in the home and having a humanoid in every home, et cetera. But the flip side of that is, you know, having these kinds of humanoids or any kind of AI embedded in the infrastructure, embedded in the systems we've got, bringing together facial recognition of street protesters, identifying them, looking at their records, making an arrest, whatever it might be. And all of a sudden it's all happened before any human brain can really put any of this together.

Yeah, the real-timeness of that, the fact that these systems can first of all sense information and then process it, and the context therefore of the application of AI AI, it's responsive to that. I was at dinner with a friend last night who was repeating the thing that you and I have heard a whole bunch, but it was astonishing and deeply demoralizing and frightening to him. He said, do you know if you jaywalk in China, you get a text saying you've been fined $25 and it's before you hit the curb on the other side of the street? And I said, yeah, that's what's going on. And that's exactly the ambient quality of this stuff. And it's, yes, it's the programming in real time. So, you know, long ago, computer systems had to think about how to program everything. If this, then that and things. In this new era, it's really operating on real-time information, real-time insights, real-time context that is the special flavor of AI in this ambient world.

Yeah, and who does AI ultimately work for? Was a brilliant question by John Battelle. And I thought, that's one that's so easy to overlook actually, isn't it?

Yes. And it's, again, with these expert surveys and the real-time research I was doing at the Pew Research Center, one of the striking things about social media as we were studying its rise, you'll remember this, in the earliest days, 2008, '09, '10, where Facebook Facebook, Twitter, and things were just coming into being, there was almost universal hope. This was going to democratize information ecosystems. It was going to allow many more people to tell their stories. The gatekeepers were smashed, and look at what, you know, how they were limiting the capacity of people to learn about each other and learn about things. But the thing that didn't get insight at first was that those systems were optimized for attention and engagement. The business model was, you know, keep people on these systems as long as possible because we're selling ads to them or we're selling them to advertisers, really. And when you optimize— who it turns out that we learned around 2012 with Edward Snowden and then later in the sort of misinformation fake news era of the 2016-17 type— that when you optimize for attention engagement, you go to the bottom of the brainstem. You find out what angers people. You get them paying attention to your site and your information by annoying them and by creating enemies for them and creating opportunities for them to attack their enemies and sometimes doing even worse than that. The new thing that AI is adding to the mix is not only is it optimized for attention engagement, it's optimized for intimacy. If you, you think about the AI agents that are being created and the systems that are being created to encourage interactions with AI agents, is that you have to trust them with your information to act on your behalf in the world. And you have to, you know, figure out how to cope with the other systems that are representing other people and institutions in the world. And so the way to do that, of course, is to build intimacy. That's how you build trust, is you are vulnerable, you are more disclosing over time, volume of interaction increases with the people that you trust and the institutions that you trust.. And so you study the incentives and you study a lot about how systems are going to unfold. So one of the reasons that you all were arguing, you all experts were arguing for institutions, is that's just even more dangerous than just attention and engagement. If you're optimizing for, I'm going to like you, I'm going to trust you, I'm going to maybe fall in love with you. That's a handoff of human agency and information about people that's exploitable and that can be turned to great harm in the wrong hands. And so there's a— I think that's exactly what Jon Battelle is arguing for here. If you are not in charge of the agents working on your behalf, if you haven't sort of made sure that you, at the end of the day, you the person, are making the final call on things that matter to you, That's just a recipe for the loss of human agency and the loss of human capacity.

I completely agree. And I mean, when I've been thinking for a long time about, you know, the psychology of the self, the biology of the self, and then this third dimension, the technology of the self, I've always said we don't own the technology. So it's a part of ourselves that we're now connected to and dependent on that we don't actually own. I mean, you could take it to its extremes and think, you know, in future if you're uploading or downloading skills or whatever and you're using those, if you don't own the channel, you don't own the pipe, you don't own the connection, you know, you are just signed up to some terms and services. Anybody can switch that off anytime they like. And you can think about that in any number, hundreds and hundreds of different examples. But it reminds me, you talking like that reminds me about anthropomorphism, 'cause that comes out in the survey as well. I mean, people like David Gunkel, who I just think his work is fantastic. And he's been talking about robot rights, robot rights are coming for what, 15 years, probably more. And that shows up for the first time really evidentially, I think, in this report. And people seem to have a worry about robots or AI having its own rights. And at the same time, that it's going to be anthropomorphized in order to create that kind of digital intimacy that you've just been talking about. So those seem to be another two themes that come out very clearly again this year.

Yeah. The ultimate question there is, of course, when human rights clash with AI rights, how are the courts going to, and judges going to decide? You mean the AI courts? Yeah, maybe that's, yeah, that's part of the mix here. But, you know, there have been some pretty powerful arguments that we have already lost sort of the human-centeredness of the rights debate by handing so many rights to corporations. I mean, they act on our behalf, there's a logic to thinking about that, they make deals on our behalf, they regulate in some aspects of our lives with human capacity already being adjudicated, but, you know, it just creeps people out to think that these systems might be given the same deference that corporations are In America, sort of the most debatable or the thing that's most hotly debated is political speech. And now, you know, it's enshrined in our laws in America that corporations have speech rights and can give money, and money is equivalent to speech, in basically any way that they feel like it. And so we've already sort of lost the claim of human uniqueness in a legal context. And there, of course, there are other people who are arguing at some point, once we understand maybe what consciousness is and things like that, there might be ways in which these systems themselves, with all the connections they make and all of the intricacies and sophistication of them, they're eventually going to have feelings. They're going to potentially suffer. And that's a hairy one for us to be thinking about. And it's definitely in our future. Yeah.

So Lee, as we just start to conclude this, what If we missed out, what have I not asked you about that you think is really important that came out in the report and people listening to this need to know about or they should dig into a bit further?

I just want to carry on the argument about the John Patel argument basically operating on our behalf and things because one of the, to me, surprising predictions here that just makes total sense once you think about it is that we are heading into a world where we're already bound up in our relationship with AI agents in some respects. The social complexity of that is going to be pretty staggering. You know, we right now, our brains have evolved for managing relatively intimate relationships with, you know, if you believe Robin Dunbar at Oxford, for about 150 people. You know, we to know who they are, who their alliances are with, how to work with them, to figure, you know, have a theory of mind of them, all that sort of stuff. Stuff. And in the arrival of the internet, I started sort of trying to mimic his work, and it's, it's so sophisticated, it's, it's not possible to do. And it was not easy to do with the available tools. But maybe 500 to 1,000 relationships are things that we manage, even, you know, folks that we don't necessarily know. But with these tools now and connectivity that was baked into them and the social graph that was baked in them, we've got, you know, orders of magnitude more relationships that we are regularly trying to work our way through. Once you start adding AI agents to that, both my agents acting for me and your agents acting for you and corporate commercial agents acting for them and government agents and their AI systems acting for them, the complexity of that is going to be mind-bending. And again, it's sort of the call for institutions to sort of serve our purposes and to make sure that we just don't get overwhelmed by that, that we don't become inundated in ways that we can't possibly manage our ways out of, that we lose our agency almost by default.

I was gonna say exactly that. Are we even involved? Like, are we even, is the, I know it's the classic, the human in the loop, but it's difficult to know at what stage, what interjections, what permissions, for example, because when you talk about frictionless and other people did in the report as well, When I hear frictionless, a lot of the time I'm thinking permissionless. It's just, it's completely permissionless. Yeah.

When we did a survey of you experts a couple of years ago asking about the prospects for ethical AI, the grand finding of that was whose ethics? 'Cause we all walk around in our heads with different moral values, with different contexts of our moral upbringing, with different schooling in and around how to be in the world and how to think about the world, how to be a good or intelligent agent in the world. And so customizing that for the variety of individuals that are going to want to enact their agency and their morals in circumstances where others might not want them to, that, you know, it's just, you can see how this spins off into computational problems and cognitive problems that we've never faced before.

Yeah, exactly. And to take that point in a couple of others that you've been talking about and we've been touching on. I think if you don't have your own agency, not only can you act on your own behalf, but you can't envision a future for yourself either. And I think this is really important. And I've written some papers there on ResearchGate, and I know, and other people are starting to look at this, are thinking this whole AI-mediated world feeling quite pre-decided. And pre-decided on the basis of, as you say, you know, not your ethics, my ethics, but some invisible, potentially corporate-based ethics. Who knows? Because you don't actually know, you're not in the system. The system is applied to you, but you don't necessarily have any agency or any say in it. And I think this is really interesting that without this sense of agency, regardless of thinking about what's human and what's non-human, without it, you can't create these possible futures, these plausible possible futures, a preferential future, and Actually, you're just coasting along being sort of nudged and directed in somebody's version of the future, but isn't necessarily yours. I think it's really interesting that Carissa Volese has just come out with her book "Prophecy" as well, which seems to be talking about something similar, the algorithms are pushing you in a certain direction. And I would say it's not just that, it's that you don't notice that and you don't even know that you could have a choice in a situation because not only is a choice not given to you, it doesn't even occur to you that you should have a choice because everything is running on these rails, if you like, these AI rails. I don't know what your thoughts are on that, whether you, I'm sure you've been thinking about it, but in this pre-decided algorithmic world, who is actually defining and deciding our future? Because it doesn't even feel like it's us.

Yeah, some of the most interesting insights from this new report was to think about not so much the enablement of the kinds of things you're describing, but how to create adversarial ways to push back against them. I mean, in some respects, sort of the grand— one of the grand insights of the Enlightenment and then, you know, the creation of the kinds of governments that we currently have is to know that there are disputes inevitable in any kind of human environment, and to think about how humans can enact their interests by contesting things, by demanding answers about who's responsible. Yeah, by even declaring if an AI system screws up, who's going to take the fall? The current argument might be, these are black boxes. We don't know. The forensics are too mysterious and we can't dissect. Well, then the argument should be, well, if you can't figure that out, maybe these systems shouldn't be given the kind of power that they're given. So it's not so much how to enable all the glorious things or how to— some of the possibilities of the dysfunctional things you're talking about, but just sort of putting in place the kind of institutional pushback and friction and giving humans the right to sort of assert themselves in environments where, as you're so lovely describing, they might not feel that they've got a choice.

Yeah. So just to wrap up, did you see the students at the university? There was a university in Florida, you'll know it better than me, who had a point of view about that speech, the graduation speech, when it was put to them that, "Oh, we're launching into the world of AI, the new era, the next industrial or intelligent revolution." And and she got booed. What were your thoughts on that?

Well, there's this really emergent conversation in the United States, I think in the UK as well, as the backlash is forming pretty aggressively. And the points of friction are, there are a couple of them. One is literally the data centers. So there's a lot of new evidence starting at the place I used to work, the Pew Research Center, but now lots of polling being done. There's a NIMBY, not in my backyard sense of this. Don't build them near me. They're not going to be good. They're going to hike my electricity rates. They're not going to create meaningful new jobs.

They make so much noise people can't even sleep.

Yeah. And at the student level, the students who booed that speaker were humanities students. So it was the people who are by definition of how they've organized their education are thinking about these existential questions. And worried about human capacity in the way that folks who study humanities think about it. I mean, that's the glory of the humanities. You think about this stuff and ponder the mysteries of it and think of all the ways it gets screwed up and all of the ways that we fight back. But that's, you know, that's their frame of reference. So having somebody come in and say, this is going to be a big force in your lives, it's not necessarily— not knowing your audience, among other things, but the ways in which the backlash is forming feel to me quite like you could begin to see it with social media around 2014-15, that the original enthusiasm was very much sobered by the Edward Snowden, the information wars, and all that stuff of the 20-teens, and that infrastructure of opposition and skepticism and pushback and where wariness is now in place thanks to social media. There, and almost immediately now, there was a community of smart people and smart institutions that basically pushed back or just sort of say, wait a minute, let us test it before you let it loose on the wider world, or here are the things that you, uh, you're going to get sued for if you don't do it the right way. And so that's now, you know, already in place in the United States. And this backlash, particularly among young who are deeply worried about their futures, not just humanities students. There's this dread coming out. Remember their education, the dominant force of their early education was COVID and getting remote learning and all of the trials and troubles of that and losing your friendships, losing your extracurricular activities, not being able to make new friends, not being able to socialize. That's all informing of sort of deep skepticism among the young in some cases for the future they're heading into.

Well, that seems a good note to end on, but let me just give you the opportunity to say anything else you'd like to about the report or the importance of the main findings and also maybe hint at what's next.

I mean, people are hungry for answers or hungry for insights that can help them enact their agency. And so thinking about that, and probing some of the obvious questions that came out of this survey. You know, what is the flavor of the backlash? What are the ways in which these companies are going to have to adjust to the growing opposition that's coming? You know, for a while we said that there's an arms race between the United States and China, and that's in American policy, that's the primary thing. Now the Trump administration itself is sort of saying, wait a minute, we're watching the models, Making them report to us when they're coming into being is a good thing. So there are ways in which the exploration of human capacity and human glory and the sort of great things that have helped us, you know, for better or for worse, conquer the planet, that feels like the call of future research.

Yeah. Well, I depend on your brilliant reports. They're amongst my absolute favorites. Thank you for doing them, obviously. And they're always splendid. We'll put a link to everything obviously in the show notes. But just to thank you again, Lee, for your time. It's a privilege to get to speak to you and thank you for all your and Jana's work on this. And I hope that we can get more and more people to read them and contribute, I'm sure.

Thank you so much.

So fun to be with you, Tracey Follows. Meechine Dialogues is a special series from The Future Review.

Find us on Spotify, Apple, and YouTube, or at meechine.com. That's M-E-H-Y-P-H-E-N.com. Hine.com or at traceyfollows.com and search for podcasts. We'd love for you to leave a rating if you like the show and do subscribe so you never miss transmission.