エッジ・ダイアログ：アロンドラ・ネルソン博士と語る「未来は誰のものか」

Your film is now ready to be shown. Good morning. I'm Justin Hendricks, editor of Tech Policy Press. We publish news, analysis, and perspectives on issues at the intersection of tech and democracy. Today's episode is the first in what will be a bimonthly series with the Privacy Center at Georgetown Law School that we're calling The Edge Dialogues. The conversation you'll hear today was recorded during a virtual event that took place on October 22nd, where I was joined by my co-host for the series, Privacy Center Executive Director Emily Tucker, and our first guest, Dr. Alondra Nelson. Let's jump right in. I want to welcome all of you here. Very grateful to everyone who's taken the time to join today.

My name is Justin Hendricks. I'm CEO and editor of Tech Policy Press. And this is the first installment of something we call the Edge Dialogues. This is a series that is designed to bring together thought leaders, practitioners to explore fundamental policy questions. We're often told we don't have the right to ask and to help us try to understand why so many years of conversations about problems around tech haven't led to the kinds of laws and policies that actually keep tech companies in check and deliver us to a better future. We're hoping not to just talk, though, about specific policy matters, not just specific questions around legislation or, you know, this policy consideration or that, but to talk about the ideas and issues that are upstream of those questions. We want to talk about our labor, war, justice, privacy, art, community, all sorts of issues that inform the opportunity space around

policy, the ideas that shape that space. And that's what I think we're going to get into today. This is very much aligned with what we're trying to do at Tech Policy Press generally with our mission, which is to provoke debate and discussion on issues and ideas at the intersection of tech in Democracy. And for this series, we are paired with, equally partnered with the Privacy Center at Georgetown Law School. And I want to introduce Emily Tucker, who is the director of the Privacy Center. And Emily, perhaps you can say a word about it and its mission. Hi, everyone. I'm Emily. I run the Privacy Center at Georgetown. The Privacy Center is focused on investigating the way that government and corporate surveillance affects historically marginalized communities in the U.S. and globally. And we are really excited about this series because it is our sense that, you know, big questions like the question we're taking up today about the future are not just left out of tech policy conversations, but are pushed out of conversations by those who would prefer to talk

about technology. And so it feels it's like an intentionally adversarial move to bring these things into the center and to push back against the whole framework of the conversation. So we're We're really excited, really excited to see how many folks signed up for today and hope you'll come back for the next one. You'll be hearing from Emily again a little later in the conversation. She'll be my effectively kind of co-host conversation partner through this. The way this will work, we're going to first hear a set of remarks from Dr. Alondra Nelson, who is our guest today, and then we'll engage her in discussion. So I would just take this opportunity first to thank Dr. Nelson for joining us and just really to say how excited we are to have her as the very first person to join this Edge Dialogue series. She is the Harold F. Linder Professor of Social Science and founding director of the Science, Technology, and Social Values Lab at the Institute for Advanced Study, where she's been on the faculty since 2019.

That lab examines emerging scientific and technological phenomena where they meet the frustration and fulfillment of civil, political, and social rights. She's working on projects around political economy of AI, democratic governance of science and research, cumulative impact of algorithmic systems. She was previously a professor of sociology at Columbia where she established the Division of Social Science, was its founding dean. She served as the 14th president and CEO of the Social Science Research Council. And between 2021 and 2023 was deputy assistant to the president and acting director and principal deputy director for science and society at the White House Office of Science and Technology Policy. And while serving that role, was an architect of the White House blueprint for an AI Bill of Rights, which I can tell you, having covered its release and various debates around it after the fact, has been incredibly influential on AI policy around the world.

Nature named her to a global list of 10 people who shaped science. In 2023, she was included on an inaugural Time 100 list of most influential people in AI, was appointed to the United Nations High Level Advisory Board on Artificial Intelligence. And so I couldn't think of a better person to help to usher in this series and to give us these opening remarks. So, Dr. Nelson, I will turn it over to you. Thank you so much, Justin. Thanks to you and Emily for devising of this new kind of conversation. Thank you for that really kind introduction, which always sounds better with your like wonderful, you know, southern accent. It makes it sound much more impressive than it actually is. So I'm going to say a bit about futures and about who gets to build them, I think, is the question that faces us. And I think also the challenge that we face. I want to say at the top that the remarks that I'm sharing reflect lots of ongoing conversations. My lab is home to several different work streams.

I have a lot of incredible collaborators, but in particular, my thinking on some of what I'm going to talk about today has been shaped quite a bit by conversations with Maura Weigel, Karina Albrecht, and Yolanda Yu-Jing Zhang in our collaboration, which is called CINI, Sociotechnical Imaginaries and Narratives of AI. I want to begin with a type of document that I've been reading closely as part of this collaboration, my part of the Sinai project over the last 18 months, which is the tech or the AI manifesto. I want to take us to the website, log on now if you want, of the venture capital firm, the Founders Fund, which Peter Thiel helped to launch and which has on its, you know, it has everything you might think a sort of corporate organization's website might have about, you know, mission, etc. But it also includes a section that's called Manifesto. And the manifesto has this kind of wistful, interesting title, What Happened to the Future, as posed as a question.

The manifesto, the Futures Founders Fund manifesto, opens with a kind of mood of mourning, as a lot of manifestos do. Manifestos, I think, in general, across the genre, are about a kind of mood of mourning. Where did things go wrong and how might they go better? For Founders Fund, this manifesto recalls something called the Ford Nucleon, which was a 1958 concept car that was meant to run on a small nuclear reactor. Then it recalls 10 years later, the author Arthur C. Clarke imagining in 1968 that we would soon be booking commercial flights into space. And all of those dreams seemed reasonable at the time, the Founders Fund insists, but that somewhere along the way we lost our nerve. And the reason that we don't have a car that's running on a nuclear reactor right now is that we weren't bold enough, we weren't brazen enough, we lost our way. And then there is a quite memorable line that some of you might know or remember if you've read the manifesto or if you're reading along as I'm talking about it now.

This is a quote. Instead of Captain Kirk in the USS Enterprise, we got the Priceline Negotiator and a cheap flight to Cabo. It's a good line. It's a memorable line. You know, it's a line that always makes me laugh even when I go back to revisit it. But I think it's become, for me, increasingly more revealing about how this sort of genealogy of our thinking about the future kind of sits with what the firm explains as its two, quote, two primary and related interests. The first is supporting technical technological development. The second is, quote, earning outstanding returns for our investors, close quote, quote. And when the manifesto finally gets around to answering the question in its title, what happened to the future, the answer turns out to be financial. So what the manifesto sort of its diagnosis is that since 1999, venture capital as an industry has been losing money for its investors.

And so the job, obviously, of other VCs and of the Founders Fund is to no longer do that, to get out of that sort of trajectory. So part of this manifesto and this vision of the future is that the future went missing, right? The future that we imagined, which was going to be not only nuclear reactor, you know, fueled cars, but also outstanding returns. We sort of lost our way. And how do we get back to that way? So part of the answer to that question becomes, I want to suggest, sort of creating and articulating other kinds of futures. Like that future witnessing, and we want another kind of future, when which the portfolio never stops performing and which technological development sort of moves ahead, undisturbed, unmitigated, and with the expectation that technology itself is supposed to have some shared good. I wanted to start here because Justin and Emily and their capacity of sort of framing up the work of the Edge Dialogues as they're getting underway have asked kind of a couple of important questions.

I think who gets to define the future and who does that future, whom does that future serve? So what strikes me in particular about the Founders Fund manifesto and AI manifestos, as I'm sort of describing and writing about in my academic work more generally, and true to some degree of all manifestos, is that the future it describes is something that you can own a share of, right, in a kind of capitalist sense. And the people who are describing it, the VCs in this case, are the ones who hold the shares, right? It's not understood necessarily as something that we build together. So I want to think about two things, and before we move to conversation, I want to look at, think about closely the stories that Silicon Valley tells about the future, especially in the form of the manifesto, and why those stories matter even when their predictions fail. And then I want to think about a different way that societies have tried to imagine

and in forms of governance, what comes next, which is through policy, and suggests that policy, if we're kind of unleashed from thinking about it in a kind of strict and wonkish way, leaves lots of room for more of us to think together and collaboratively about what a future might look like. So I know that Justin and Emily, they have announced in the Edge Dialogues that they don't want to talk about AI governance. So I find that refreshing, even after years spent in that space. What I want to say, so I'd pivot slightly and say that this is not a talk about AI governance per se, but a talk about power, about the power of futures, about the power of narratives about the future, and about the increasingly small number of people who are laying claim to the future and what the future looks like, what its shape of it, you know, how we might imagine it before the rest of us have had a chance to have a say. So manifestos are much older than Silicon Valley, as you are well aware. And you can think back to, you know, you can think back to the early 20th century in a small town in upstate New York with Elizabeth Cady Stanton and women and men when gathered at Seneca Falls to adopt a Declaration of Sentiments.

It borrowed words from the Declaration of Independence and changed them to read all men and all women are created equal. What that manifesto shared was that their authors had almost no formal power. None of the women at Seneca Falls could vote, but they decided and articulated that they had a voice. I've been keenly interested in how manifestos, the tradition of them up until fairly recently, have come from people outside of the system or people who've been trying to sort of change how we think about the system or get inside of the system. Often, authors of manifestos lacked power. So even if you think about the well-known Italian futurist manifesto of the early 20th century, that would make fairly misogynist claims about women and their relationship to technology, would be anticipating and praising kind of fascist turn in politics. But it was very much the sort of sense that they didn't have power as artists, and I think that was also empirically true. And manifestos attempt have historically attempted to make up for the lack of force or the lack of power through language,

through powerful language, compelling language, poetic language. The manifesto is powerful as a genre, both past and present as an AI manifesto, because it doesn't negotiate. It just declares. It says out loud what the authors sort of intend for the future to look like. So the manifestos of the AI era include ones that I'm looking at closely. Mark Andreson's Techno Optimist Manifesto. We might also consider Sam Altman's essay, The Intelligence Age, Dario Amadei's essay, The Machine of Loving Grace, a statement by Palantir that was posted this spring. It was a distillation of Alex Karp's book laying out its vision of a technological republic that was also taken up as a manifesto. So the AI manifesto is becoming very familiar, but the authors are new in some ways, right? these are not outsiders. These are insiders. These are incumbents. These are very powerful people who command not only capital, not only computing power, not only, you know, have the levers of the

global political economy, not only have the ear of presidents, but they also write now with the voice of the insurgent, the voice of the outsider. Andrundson lists among his enemies the precautionary principle. He writes that any deceleration of AI will cost lives and that the deaths in an AI system might have prevented how to have been built faster amount to a form of murder. So the language of insurgency, of being an outsider now comes to from the barricade of social movements, of artist movements to the venture fund. So you might reasonably wonder whether any of this matters since our future visions fail all the time. We're terrible at predictions. But I want to bring to the conversation as I'm wrapping up the work of the sociologist Yance Beckert, who gives us a useful way for thinking about the relationships between futures and markets. Anyone who puts money into something new, Beckert points out, faces a basic problem. The outcome of that investment can't be known in

advance, right? So you can't, the spreadsheet, even if it's well simulated, well modeled, can't tell you whether or not a product or technology is going to pay off in 10 years. So he suggests that economic life, and I think particularly it's true of this moment, kind of runs, is fueled in part by fictional expectations, by stories about futures that are credible enough to get people to commit to the hype, commit money, commit effort. If we're consumers, commit to using the tools, even if they're often in failure mode or quite fallible. The story doesn't have to be true or accurate to work, right? The story of the future only has to be believed by enough of us to sort of make it true in some other sense. So I started thinking about AI manifestos as these kind of expectation devices following from Becker. It's as a tool for making one particular vision of the future feel credible enough for some of us to act on.

The old historic manifesto used dramatic language to make up for authors' lack of power. Today's AI manifesto uses dramatic language to make up often for a lack of evidence or for a lack of demonstration and market value, speculation in the market rather than demonstrated value. The authors of the AI manifesto, I think, have nearly everything to prove except proof that the vision will actually arrive. So we're all kind of in this state of expectation. There's another way that we can think about futures, and I want to suggest, I think somewhat probably counterintuitively, that policy is one way to do that. We think that we tend to think of policy as the opposite of imagination, as the opposite of a future that is compelling, as a future that we want to share, as the brake on a car, red tape that makes us angry. But I think if you look closely at almost any policy you will find a future a picture of a future that someone wants Right So I like to understand and define policy as an articulation of a desired future state So it means that taking something as ordinary as like a regulatory code or a zoning code

you might think that when a city requires sidewalks or buildings that mix shops with apartments, it's not only just creating a policy or a rule or regulation, it's also describing a future in which people walk to a corner store, they run into their neighbors along the way, that children are safer on the streets because there's fewer cars. Or we might even think about the rules that families make or the policies that we make and the ways that we make our families. When a parent says there can be no screens after a certain time of day, it's in part because they're picturing a future for their children who sleep through the night without blue light disturbing their rest, but who can also sit with a book or an article and sort of imagine that kind of future of intellectual life and engagement and discovery as well. So I want to, so each of these sort of policies, you know, city policy, municipal policy, federal policy, or the policy of a home is itself a kind of small forecast and that has people attached to it

who are investing and how it turns out. So I think what sets public policy apart as a future from a manifesto narrative is everything that sort of sits around it. Policy, you know, often has to build sort of public consensus, right? There's often comment periods. There might be hearings. People are weighing in. It invites the back and forth that the manifesto shuts down. It can be revised and over time. And I'm increasingly wanting us to take a page from software versioning and think about even government policy as having a kind of iterative life if it might be a version one or a version two, to sort of be honest about the sort of the fact that the work of policymaking and of envisioning and crafting futures is never quite over. So I want to just just in time we mentioned the blueprint for an AI Bill of Rights. I believe if my math is correct, that today is actually the fourth anniversary of the issue of the White House blueprint for an AI Bill of Rights.

So I wanted to lift that up as a way of kind of thinking about policymaking, a policy white paper in this case as a kind of future. You know, it set out five principles that automated systems should be safe and effective, a condition, that benchmark that we are still trying to meet, that the news reminds us every day that we've not met, that they shouldn't be biased against us, that they should protect our data privacy, that you should be told when one when AI is being used for consequential decision, and that you should be able to opt out and reach a human being if something goes wrong. So the grammar of those sentences are about a future, about a desired future. They speak not only to investors and venture capitalists, but also to a broader public, both about what they might imagine and about expectations they can have about what they are owed about AI systems. And, you know, I want to just say as a sidebar, because it's, I think, the anniversary, we're coming close to the anniversary of the blueprint for an AI Bill of Rights, that, you know, what we didn't fully anticipate is that it's had significant kind of afterlives.

Justin sort of suggested this. It's been, you know, it's inspired educators. Governor Ron DeSantis of Florida has stood up his own AI Bill of Rights, which draws on some of the five principles that I've mentioned to you, that it's had sort of many, many lives because futures, when they are not just about a few of us, right, they can sort of travel and more of us as in the sort of older sense of the manifesto can imagine themselves as being part of the imaginary or the narrative for the future that's being set forth. So Justin and Emily asked what a future shaped by people and communities might look like. And I hope that maybe sort of spreading your aperture, opening the aperture of what you think even policy is, that we can begin to realize that we have more examples than we sometimes realize. It's sometimes difficult in the context of, I think, the chaotic space of how AI is being sort of diffused through our lives right now, often without our consent to imagine otherwise, to imagine a future that we indeed shape.

So what has happened to the future? To return to that question I began with, I want to offer a different answer from the Founders Fund. The future there morning was always at the bottom, a projection of investment returns. So it's no surprise that it showed up first on a balance sheet while other futures were being written all along, I think, in homes and union halls, at city councils and state legislatures, and kitchen table conversations about screen time, about whether or not AI should be in schools and more. The manifesto began as a voice of people without power who were speaking into and for and with and about their futures. Its migration into the hands of some of the most powerful people of our time tells us something about its power as a genre, about the power of sort of speaking into the world about the future. But we also have to sort of reclaim that power as something that belongs to all of us. The AI manifesto can be reimagined as a place of struggle where some of the most consequential sort of imaginaries and frames about what our future with these new technologies might look like if we're allowed to play a role in it.

each of us has a claim on the future, and I want in this conversation for us to begin to think together about what that looks like. Thanks. Thank you very much, and what a great start to this series. And my immediate takeaway, of course, is I'm going to have to go and explain to my children that the policies of our household are my way of, you know, describing their desired future. Yes. Absolutely. But I'm going to turn it over to Emily to offer her kind of first reaction to your remarks. Thank you so much, Justin, and thank you, Dr. Nelson. I have a sort of half-prepared riff, which I think actually turns out to dovetail really nicely with your remarks, but I'm going to kind of do some of that, improvise a little bit. But I just wanted to encourage those in the audience. If you have questions, we are going to have a pretty robust Q&A later, so please feel free to start entering your questions now as they arise, and we'll be looking for some good ones to pull out. So I think one of the things that feels quite eerie to me

about the particular version of the future that is in the manifestos that you describe and that I feel like increasingly pervades conversations, not only about technology, but about all kinds of things in politics and culture, is what looks to me like a kind of imaginative impotence to it. So the rhetoric around AI is so sensational. It's Hollywood blockbuster promotional rhetoric, transformation, revolution, human extinction. This is the kind of language that's used to talk about the future that we're supposedly entering. But if you look like you tune into the actual details of the world that is envisioned by the extremely wealthy men who run the tech industry, what they add up to is something which strikes me as actually quite fallow and sort of disappointingly retrograde. Like even just the surface aesthetics are things like flying cars and moon colonies, which, you know, while I will not deny my own nostalgia for certain 1960s sci-fi literature and certain 1980s cartoons, those are things that are very much striking me as a past notion of the future.

And the ideas that underlie the aesthetics are even more clearly not new, but are in fact a way of doubling down so hard on old ideas that they become kind of flat and lifeless. So like even just at the formal level, what an LLM does is to endlessly recombine data about the past to make probabilistic calculations about the likelihood of one word following another in the world that gave rise to the data set. So it's not even really a calculation about what's likely in the future. It's a calculation about what will seem likely and familiar to someone right now when that's someone's own recent past is of the average kind. and this just seems super important to me in light of what you were saying Dr. Nelson because what real transformation, what real revolution, what real new ideas requires is the opposite of probabilistic calculation based on past data. It requires us to be able to ask what if our past ways of understanding the world were just totally wrong not just imprecise or biased but fundamentally confused and that's how people who make real breakthroughs actually think.

So like take Einstein's special theory of relativity, for example. The problem that he was confronting was how do you explain the fact that the speed of light remains constant no matter the apparent velocity of its source? And his solution involves showing that our prior understanding of time and space themselves was mistaken. So the length of a body and the time relation between two events are not fixed quantities that are the same for all observers. space and time are not two separate and independent realities, but are different aspects of unified space-time. And that discovery within the existing frameworks of thinking about these questions at the time was infinitely improbable. There was no probabilistic calculation based on past data that could ever generate it, because the statement of the theory in the old system would sound something like, it was brilliant, and the slimy toes did gyre and gimble in the wave. So Einstein theorizing didn't involve probabilistic reasoning about what seemed plausible to the science of its time, but a fundamental rethinking of its most basic categories. And so it seemed really clear to me that at a moment when so many of our institutions and so

many of our normative frameworks are falling apart in a pretty spectacular way, that is what we need, that kind of fundamental rethinking of basic categories. That's what we desperately need. And that means we need human beings who are capable of doing that kind of thinking. And that's not the kind of thinking that you're going to learn how to do through interactions with the chatbot, for example. You know, my students come to me and they'll say things like, I feel like the chatbot is stealing my education, but everyone says, if I don't learn how to use it, I'll be left behind. And I just find myself thinking, what is the future going to be if a whole generation of young people decide that it's just too risky for them to try to undertake the sort of intellectual and creative work that's actually necessary for human beings to think something new? And it seems to me like persuading people that this is in fact too risky is part of the political strategy for the big tech companies. They already have such a disproportionate hold on global wealth and power that a future filled with free thinkers who may be capable as individuals or more scarily to them as communities of pushing past existing frameworks.

That's kind of an intolerable existential risk for them. That's like an uncertainty that they can't afford. And so, you know, there's a kind of marketing by gaslighting that's going on where the industry is feeding us a rehash of sale ideas and telling us we're on the cutting edge. And to make yourself believe that entails a kind of soul-destroying nihilism. But the alternative is to realize that the people driving the bus are headed off the cliff and that it's up to you to stop them, which is really scary. And so that just seems to me like something that is hard to policy our way out of for so many reasons. but one of which is that policymaking processes tend to be very captured or to a large degree captured by many of the same people and interests that have co-opted the manifesto, the communicative device. And so I guess I'm just wondering, you know, for policymaking to be alive in the way that you are very compellingly suggesting that it could be, people need to have an idea of themselves as powerful and that idea has to have some reality around it.

And, you know, even really strong policymaking, like policymaking processes that happen at the municipal level tend to be more inclusive. They tend to be at a space where it's more possible to show up as just like an average community member who's trying to think about the future. But even in those spaces, it's more like there's a policymaking space that's kind of externally curated and the community is invited. And it's not like those deliberative activities are arising from the political community. And so it seems to me like there is some kind of fundamental shift that's needed. And so I guess I'm having just listened to your really compelling provocation. Like, do you think that fundamental shift, is that a fundamental shift we should look for in policymaking, like in how it's constructed? Or is it something that has to come prior? I guess I would say, you know, I also I offer policy as a kind of future making because it is the most counterintuitive and sort of extreme. And I think there's lots of other ways in between, between a tech billionaire's manifesto and policymaking that we can imagine sort of making future.

So the sort of the counterintuitiveness of the policy, policy is future is intentional, I think, to sort of, I think, elicit a little bit of exactly some of the questions that you raise. I think we can, you know, certainly there is a way, I mean, we're even experiencing it right now with the regulatory capture that's happening around AI, federal policy, AI policy. So it is always going to be the case, I think, particularly when you're dealing with science and technology, that you are going to have these big incumbent, powerful figures there. And so, you know, the question is, in what administrations or under what types of government can there be different kinds of voice for others to articulate what they want their futures to be like? So, you know, we might recall it was the Obama administration and partly with some of their science and technology policy that began to introduce sort of challenges. And we sometimes will think about the challenge that we had for the electric vehicle that sort of brought, for good and for not, Tesla into our lives.

But there were also challenges around ideational challenges, like what are other ways that government might do its work? So there's just other ways to bring people into the work. And I think that, you know, certainly the way the table is set in federal policymaking, I think, makes the bubbling up of, you know, you might say sort of more minority or marginal futures much more difficult to arise. But I also think we're seeing with the example of the local politics. I mean, we can use two examples, actually, of politics over the last 25 years around school boards at the state and local level. Right. And people can agree or disagree with what those futures were around schools and charter schools and the pride of his privatization of education. But that was the articulation of a different kind of future for education in the United States being manifest at the municipal level. Closer to the conversation for today would be the politics around data centers and around a different that is very much a different articulation of a future for AI. In fact, what's been so, I think, interesting and compelling about that space is that I think it began very narrowly with, in some instances, narrowly with people living nearby data centers, people having very proximate concerns about noise, about water, about energy, to becoming, you know, a kind of big tent, you know, sort of social movement that is also around power and inequality, around the use of AI more generally.

I'm sure you have people who are at these protests now. They're concerned about the mental health of young people and about what chatbots are, you know, the kind of delusional facet of chatbots. And so, you know, I think folks have talked about it more simply as sort of like it's the physical thing. AI is not physical and data centers are physical thing and people mobilize. And I think that's true. There's something about the grounded politics of data centers. But I think it's also true that these have been places where you can come and speak to public. And I think, Emily, where I would disagree a little bit, you were saying that they didn't, you know, these were kind of these local instantiations that didn't emerge. But I think when you have a data center social movement that is emerging out of other different kinds of community conversations and concerns, out of other school groups, church groups, other kinds of organizations, that I think we can imagine them as having kind of real local roots in a way that's important and that matters. I'll just pick up on that and just say, you know, you mentioned my southern accent, the place I acquired that in rural Virginia.

One of the early data center fights that I paid attention to was in the kind of rural community that I grew up in. And the thing that most amazed me about it was seeing pictures in the newspaper of, you know, men in MAGA hats next to folks who, you know, had opposed the pipeline, you know, had opposed things on various environmental grounds. And this was in the height of the kind of 2024 election cycle and just after it. So people were effectively polarized across the board on almost everything, and yet that seemed to bring people together. And I think I've really heard so much of what you're saying across the course of so many conversations I've had over the last few years. I'm thinking about activists in Tucson who opposed a big data center there and who, in many cases, were using that opposition, the conversations in the community, You know, of course, starting with concerns over energy prices and consumer electric bills and water, which is a major issue there and fear among many in the community.

But then beginning to branch out from there and ask questions about what kind of community economy do we want to have here You know what sort of businesses do we want to occupy this space that we call home Where do we want to work What kind of jobs do we want to have And I think that you know, those types of conversations appear to now be happening across the country. And I'm beginning to see, you know, some signs that, as you say, it's spreading out to the question about AI in schools. And that seems to be another one that's coming kind of fast and furious right behind the data center concerns. It's spread out, of course, the concerns around surveillance and lock cameras in particular. So I guess I'll ask you if you see any other prospects for kind of joining up some of this popular consideration concern and allowing it to kind of create a space for different features to imagine. What would you recommend there? I think, you know, the flock

example is a really excellent example and I will await the actual research of like scholars to sort of help us know what really happened there but you know in sociology there's this idea of social movement spillover that you have you know there are people or ideas or things that sort of have this kind of Venn diagram effect that catalyze action and movement in other places and I suspect that some of what happened with the flat cameras in addition to right already having long-standing existing and some communities resistance to forms of surveillance. Flock cameras, only one among many that one, you know, might call out and name in various communities. I'm thinking also of things like inaccurate AI surveillance systems like ShotSpotter that are sort of, you know, listening to potential violence in neighborhoods. There's been a lot of, I know there's been mobilization in Chicago, for example, around city procurement for those kinds of tools and technologies and the implications of those for communities. So I think, I mean, it's a particularly hard time for that question, given all of the

polarization. But I think the image that you gave us, Justin, of kind of left activists and folks in MAGA hat sort of being aligned around this, I think is, I think, exactly what we might hope for. But I also do think that the politics around AI are distinctive in some ways. I mean, they come a decade hence, you know, a decade, 15 years, 12 years after we have been really unable to get a handle on social media, right? So there are people who carry, I think, frustrations around our inability to govern or make social media be sort of healthy for society, even as it's making lots of money for lots of people. And I think it is an issue that has been shockingly, I can't think of another issue that's just been in both the academic polling and the political polling have continued to increase the negative sentiment. And so that's actually quite a shocking thing, and that increase in negative sentiment is this growing kind of cross-partisan, I think, concern

that is leading to this widening set of issues and concerns being put under kind of data center activism. But I think it's surely not only schools, it's people worried about their jobs, it's folks worried about, as I mentioned, the safety of young people, it's people worried about whether they can get, you know, a fair shake to apply for a job when you've got all of these algorithmic systems. So, you know, I think that we are getting slowly a vocabulary and sort of local mechanisms that I hope will have some national impact to sort of speak to, I think, the broad spectrum of things that, you know, on a spectrum from kind of concern to like, you know, deep discontent to reviled discontent around the broader transformations that are happening through the expansive use of AI tools and systems. I think I'm going to already begin to bring in some of the very smart questions that are in the Q&A There are many there that I think we've already hit on in some way

Or we've begun to kind of pull threads in that direction But there's one in particular around the kind of Gender aspect of this, the kind of hyper-masculinity In the rhetoric of AI at the moment Also, of course, in the kind of nationalist rhetoric that we hear How do you think of that more specifically? I know you've kind of alluded to it a bit, but what is there in your mind about the kind of like gender aspect of this? I think, you know, we were that part of the certainly that what I describe as the manifesto, and I'm interested in sort of probing and teasing out the characteristics of is deeply masculinist. And, you know, probably in some varieties, if you think about, I think, a former Google worker who did a kind of early Silicon Valley, early in this sort of quadrant of time. I'm talking about, I think, James Damore, who was talking about, you know, having to, you know, his negative reactions to having, quote, unquote, woke colleagues in Silicon Valley.

So, yeah, I think sort of masculist to sort of misogynist, I think, you know, certainly. And I think that there was an earlier moment in which, you know, I'm particularly thinking about the 1980s and 90s, but even actually the Italian Futures Manifesto, which in some ways is kind of the uratext of thinking about a manifesto of the future. I mean, soon after that came out, there was a manifesto of women artists that responded to it. Right. And so you can imagine that this has been a conversation about technology and futures that has had sort of multifaceted voices that I think is that part of the cooptation of futures through the generic form of the manifesto that I'm trying to, I think, sort of capture in some of the work and thinking that I'm doing, both in scholarly space and policy spaces is the reduction of those who can speak, those who can build, those who can shape the future to, you know, to men who work in venture capital, to men who make technology in a certainly kind of way or deploy it and who can make decisions about both design and deployment.

So I would just mostly agree, but add that elaboration. I would just add that the masculinist impulse behind a lot of this, I think connects back to the thing I was saying at the beginning of my response about how retrograde a lot of the visions of the future are. And I think that that's not an accident. I think there actually is a longing to go back to, you know, to times where certain kinds of oppressive systems were fundamental to how everything worked. And so I think there are elements of the masculinism that's driving a lot of the AI futures rhetoric that are kind of new and different. But I think there's kind of a core of it that is just appealing to a certain backward looking. I don't know if folks have seen Jill Lepore's really awesome riff on how the tech bros are really bad readers of sci-fi. But she she talks about how, you know, the way that they read sci fi is actually it's actually about a closing down of possibility and not an opening up of possibility.

And so I see that the way that they view gender, the idea of gender roles that they're promoting is connected to that, too. Yeah, I think that's such a good point that you make and agree with that. And I would also say, given the generation of, you know, people like Teal and Andresen, which I kind of share demographically, it's also about a boyhood imaginary, right? Like, it's not just a – it's now a man's imaginary, but it was a boy-child imaginary of the future. Yeah. I want to maybe pick up on something that I think, you know, Emily was getting on to and surfaced in the Q&A as well about the conditions necessary to permit policy to be more responsive to democratic pressure. Maybe just press you on that a little bit. What do you think needs to happen in all those various levels of government that you described to produce a result that's more responsive to the individual desires, hopes, dreams for the future? You know, I think that you see it.

I mean, it has to be public contestation, right? So certainly the data center mobilizations are an example. I mean, we went from having executives of states, you know, Maryland, New Jersey, et cetera, who were great, Pennsylvania, great boosters of data centers to sort of completely rolling that back. I think even in the United States at the executive level, we've had we've gone from a language of, you know, that I've written about a little bit of deregulation, which wasn't actually deregulation. It was actually, I think, the U.S. federal government sort of quite strongly shaping AI policy through things like, you know, investments, gold and shares, taking portions of companies through the way that we're shaping immigration policy, through the way that we're shaping trade policy, you know, who can have semiconductors or access to them and not. So what was called deregulatory, I think, has been quite strongly steered and controlled in a particular way, even if it wasn't formal regulation in a way.

And so I think I would say on the democratic and more sort of people power side, that I think part of what needs to happen is an expansion of where people think that they can plug in or where they can be effectual. So if you like the sort of analog I'm trying to make is that if we think that the only way that the federal government or the executive branch or Congress can do something around AI is through making new laws, then we have not been paying attention to all of the ways that this government has been closely shaping AI futures. Right. So I think similarly on the side of people outside of government, there's a you know, there's a cluster, I think, of things that I think the data center movement broadly, again, I think, suggests, which is sort of working at the local and state level. I think even that local data movement, data center sort of mobilizations variously, and they're quite variegated, have put pressure not only just on, you know, a town or a village that's thinking about a data center, but as I was trying to say, on how states are doing technology policy and AI policy and arguably how the federal government, if not how it's doing policy, has come to think about how to talk about it, right?

So over the last several months, over the last quarter, if not two, it's become, I think, even at a federal level, very difficult to say that you don't care about the affordability of electricity and that you don't care about people's concerns about sort of the pollution, about AI in their neighborhoods. And so what, you know, so there's this kind of transom, I think, of things going back and forth. And I think what I would reject from the question is that nothing is happening and no movement is happening into reaction to what's happening on the ground. In a place like New York City, even, I mean, we've just had a there's just been a almost day long sort of set of hearings about AI by the entire city council, which sort of had lots of people speaking in lots of different ways. So I do think it's a different moment. And whether or not, you know, it is the case that someone in rural Virginia doesn't have formal levers to change policy in Washington, D.C., but there is, but certainly what's been happening in Virginia, which is the state I think in the United States with the most data centers in the country, is impacting federal policy, federal norms, the way that we talk about AI and the benefits and risks.

of AI. So I think I want to loosen up a suggestion that the only way that public engagement matters is through formal mechanisms, right? And that we are seeing that there's lots of different other ways that it matters as well. I would plus a million to that. You know, I think that people are very pessimistic about participating in movement spaces or participating in civil society democratic work because they think they look at what policymaking is and they think, oh, there's no possibility for what I want to show up here. So what's the point? But actually, like, what is most important about that work is not the outcome, but it's actually the political community comes to be in the process of the coming together to coming together to try to imagine what change could be possible. And so, I mean, one thing that I am always thinking about is how can like the possibility of policymaking is a great carrot. Right. It's a great we have a we have an idea of a change we want to create. And that's the organizing vehicle. But anybody that's ever done political organizing can tell you that like you never what happens is never what you expect.

And a lot of times the thing that you most the thing that you most want turns out to be the wrong thing to want it. And you that work of discovering that together with other people that you're doing that that you're doing that imagining with is where the power is. It's not in like, you know, I have an idea. What's my what percentage estimate do I have that I can get that idea enacted? Right. And so I guess that, you know, we do a lot of work with CASA, an immigrant rights organization based in Maryland, and they've just done such an incredible job of taking the mass surveillance environments that their communities are forced to live in and taking that as an occasion for organizing, as an occasion for building their own capacity, their own knowledge about this set of issues, and then bringing that knowledge into the demands they were already making to protect themselves, to protect their communities, to fight back against detention and deportation. and that even though the wins that they've had just legislatively might look thin on paper, that power that's been built, that community that's been built, is going to have a life that we can't predict. I think that's exactly right in the sort of, you know,

history of, I think, social movements in the U.S. and abroad proves it out. And, you know, also I'm thinking you're making me think, of course, about Theta Scotch-Pulse work on just like the importance of associational organizations, right, that sometimes they're laser-targeted on particular missions or goals And sometimes it's the association itself that sort of builds the capacity to do exactly what you're suggesting. Harihan's work is very good on this point as well. There are a couple of questions in the Q&A that kind of maybe ask us to broaden out to a more global perspective, a question around Europe, a question around India, Nigeria, maybe just zooming out from rural Virginia out to the global kind of purview And maybe thinking about your experience, you know, working on a policy at the level of the United Nations, for instance. I mean, it does seem like one of the big disconnects right now for so many people across the world, particularly when it comes to contesting the futures that Silicon Valley is making most salient with all the billions spent,

is that people don't have any direct constituency to affect what happens in those boardrooms. They aren't voting for a government that seems to have, you know, the ability to get its hands around leaders there. How do you think about that, that kind of disconnect? What's the role of global governance in that context? You know, I think global, you know, it's the best of times and the worst of times, I think. I mean, it's a very challenging time. So I was just at the U.N. General Assembly going to meetings and conversations and side events there. And I think one of the things I found most interesting and striking, so I had been on the UN high-level advisory body on AI, and we issued a report just before the UNGA that took place in 2024. And, you know, what was interesting is that there were some folks in the kind of UN blob, distinct from the DC blob, who were, I think, not really getting it. They didn't really understand why we were going on about AI.

You know, credit to the Secretary General for putting it on the agenda. But there were lots of folks who were just like, why do we care about this and why does it matter, et cetera. So split screen to 2026, where you have during it, you know, a U.N. Security Council briefing from tech executives. You have, you know, the French as part of their membership on the U.N. Security Council bringing AI to the table. We know that, you know, I know that the Democratic Republic of Congo, which has a lot of critical minerals and, you know, lots of negotiations with the United States on that, are now also on the Security Council and are trying to make part of their sort of leadership and the role that they can have in agenda setting to the extent that non-permanent members can agenda set at the U.N. Security Council. a conversation about natural resources and about critical minerals as part of a conversation about peace and security. So then there were lots of other conversations about AI from different sort of directorates, communities, missions that weren't taking place two years ago.

On one level, I think, Justin, it was the most it was like it was finally that the UNGA AI. Right. There were like there was just like the AI UNGA rather. There was just lots of conversation happening, a sense of urgency in a lot of different ways. On the other hand, I think that we are probably farther away from any kind of international multilateral kind of governance regime. And part because, you know, the United States is that has shown and the Trump administration has shown that it's, you know, kind of steadily against it. And so there's not that a will to sort of make that happen. But also I think it you know we starting to see if you look at some of the global polling around AI there this sort of variegation sort of happening So you have high negative sentiment in places like the United States but in places like Brazil and China you have positive sentiment. You have United States multinational corporations that really control

the design and deployment of most AI in a context that's, you know, that has this negative sentiment. As I hear from colleagues in the global majority, you know, there's a lot more optimism and people think that AI is going to be important and transformative for their communities in some way, even as their concerns about resources, critical minerals, water, labor, in particular, the labor that sort of underscores all of this and is little talked about and not talked about enough in the work. And so I wonder if aspirations for global governance in part have, you know, they've gone to the wayside because, you know, there's not in the sort of scheme of grand power strategy, there's not the big players, you know, are not driving that. But it's also the case that what AI means, I think, is becoming it's becoming clear that it means different things to different kinds of global members of the global community, and that their incentives and stakes and concerns and aspirations may also be different. I'm going to pull out another great question that I think is

A good one to actually pitch to Justin first. So speaking of power, how do you think media, mainstream and social shape futures, especially when it comes to amplifying or suppressing specific futures over others, considering how many tech CEOs own more and more media? One of the things I immediately thought about on seeing this question was, you know, not only this issue of media concentration and the extent to which that's an issue, But also, maybe kind of counter to the question on some level, the just avalanche of dystopian entertainment that is available right now. Everything from series like Silo on through to some of these more recent films that we're seeing, like the Musk documentary by Alex Gibney, The Social Reckoning, which is this, I suppose, slightly fictionalized account of the Facebook papers and Martin Zuckerberg and Francis Haugen and the Wall Street Journal reporter at the time, Jeff Horowitz, and that sort of set of events from a few years ago.

And then, you know, I'm also thinking about just a range of different sort of portrayals of billionaires as villains that we're beginning to see that kind of seems to be a kind of common thing. So I don't know. It seems like it cuts both ways right now. You know, there is this concentration of media power. I do think that the hype around AI very much is the salient feature of the media information environment. but there's also the kind of black mirror effect and that sort of inside of it as well, which seems to be priming people to question the future that's on offer. But Dr. Nelson, I don't know if you'd agree with that or maybe if you just have a favorite science fiction show you'd like to mention. Oh, I always have a favorite. I think, yeah, I agree with that. I mean, I think that what makes the question hard is like the sort of interaction effect, right? So that, like, you know, like social media becomes a place where people, you know, broadcast different kinds of future and about the about AI and other things.

And we're also using AI to create images and to, you know, some of the presenters are kind of AI folks. So it's a space that's both, you know, emerges out of a kind of capitalist media foundational kind of space, cornerstone that has a, you know, that is sort of woven through and through with AI and also a space that uses AI. So it's this kind of, I think, nested space of AI politics and AI futures throughout. I always understood the purpose of the media is hold power to account. And how can it hold power to account when it is itself power? I just don't see a way out of that fundamentally. Yeah, I think when we think about news and journalism, we're in a slightly different place on that question. And unfortunately, I think the dynamics being what they are, the kind of AI moment that we're in, seems like it will only weaken all of the economic incentives to produce accountability-oriented news and journalism and to send signals to publishers about the types of things that are necessary to produce.

So that's a tricky moment for folks who are in my line of work, certainly. There are so many questions that get on to smart issues here. Some of them are asking for just simply recommendations from you, Dr. Nelson, about books you find illuminating that you would recommend on this issue in the current context, on, you know, other ways that people can kind of pursue these ideas. It also, you know, one of the questions we had sort of written down in advance to try to ask you goes in this direction as well, which is, you know, who's a thinker or a writer whose work on these issues or these questions you feel like doesn't get enough attention that you feel like folks should go and pursue that maybe, you know, should get a next crack at one of these dialogues, for instance? Yeah, I mean, I think as far as reading, I'm actually reading right now. I just received it. I think you probably can't see it, but I'm reading Julian Posada's Platform Extractivism, Data Work and the People Powering Artificial Intelligence, which is this rich kind of ethnographic and analytic account of data extractivist work in Latin America that I think offers a different story.

We've got great reporting from Karen Howe and others from parts of Africa. And Julian is adding to that story. I think a lot of the extractivist work, the work about data extractivism or the technology-related extractivism in Latin America is often about critical minerals. And he really puts sort of people and their stories and data work in the center. So I think it's a really important addition to that story. I also have been reading the work of my colleague for the year, Deb Raji, who is a computer scientist, Dr. Deb Raji, who's starting on the faculty of Stanford next year, in part because I was asked by Governor Newsom to serve with three other people on the sort of expert committee to make recommendations for the implementation of his most recent executive order around AI and about the implementation. of two bills that he signed around frontier AI, how to think about independent verification organizations,

how do we get more accountability by, what were the mechanisms for doing that and how might we get more control? What are the systems of control that we need over algorithmic systems as we see agents being released or tested in ways that are not fully controlled? So I'm reading lots of Deb's work. who's she's a real pioneer in thinking about AI auditing and AI systems and going back to some of that key work. And I just have started reading a paper that came out this week from Arvind Naraghanan, Rishi Bambasani's on this paper, Syed Kapoor, and lots of other smart people about how to begin to think about evaluation and governance regime sort of after the agentic turn and to do that. And these are all researchers who think about this in a kind of sociotechnical way, like what are the use cases? So these are not just, I think, you know, merely kind of narrow technical approaches, but I think approaches to thinking about what's different about the agentic turn and how to think about that in context and the implications for that.

There's a question around kind of Vannevar Bush's NSF and that kind of vision and the vision for science and technology policy. And I'm thinking about that in particular because, you know, we just had recently a very different vision for how science and technology policy should work from the Trump White House and the OSCP. But I'm wondering, as you think about the next few years, you think about the possibility that there'll be a chance to reset science and technology policy. What do you think needs to be kind of top of mind in order to maybe open up the possibilities for the future in a different way? Yeah. So, you know, I think that we are in ways that are sort of appropriate and inappropriate ways that are within the law and potentially sort of outside of standard kind of protocol for appropriations and the like, sort of totally transforming the sort of science and research ecosystem in the United States. And, you know, I think that my scholarly work and I think my interest in policy work and, you know, the time I served in government was really always about being a reformer.

I mean, I think those are like, you know, it is a sacrifice. It's an honor, but it's also a sacrifice to work in government, for example, or do public service. And I think part of the, you know, part of the appeal is that one has an imagination for a way that we could do things better. This is currently how government's doing it, but I've got an idea or a process or, you know, we might be doing it another way. There might be a, you know, another future for how we do this particular policy. So, you know, as much as I lament how some of the changes have happened, I mean, literally through ignoring Congress's Article 2 powers, I mean, the kind of lawlessness of the change is lamentable. and at the same time I think that there were things that needed to change and would have liked them to have been done through a process that was a bit more legitimate and that included the public. That said, I think what this moment does is sort of shine a light on the fact that that moment that even with the White House's new document about a new golden age, right,

that it sort of is trying to take up from, pull a string forward from a kind of prior, it assumes a prior golden age. And I guess I would say, and this is some work I'm doing with the sociologist Jenny Reardon and others, is like that prior golden age also wasn't golden either when it came to the question that we've been asking about the inclusion of the public. So we have in our imagination of Annabelle Bush, particularly in Washington, D.C., this idea that it was like a perfect vision and we set it up. And for almost 80 years, it kind of worked perfectly. But if you go back to, you know, Bush's important report, Science, the Endless Frontier, and some of the policy work that, you know, policy frameworks and ideas that were circulating in that moment, none of them had a vision for the public at all. Like it was supposed to be this kind of trickle down effect that like, you know, let us, you know, count myself among them, you know, the us, like let we scientists and researchers do what we do. We're going to set up our own internal systems. And we promise that good things will trickle down to American society.

Right. And the public didn't have a role in setting that up. And so what I hope will happen in the coming years that is not happening, certainly with, unfortunately, with the new White House document, which I think has even retreated even more from any kind of public accountability, is that we will have a way to have, you know, what Jenny and I call a public science, one that is, you know, open. I mean, some of the work that I did in my time in Washington, D.C., was sort of how do we get taxpayers access to research that they pay for when it's published, for example? How do we have, if we're going to rebuild everything or have to reimagine, how do you do that in a way that includes explicitly public accountability and public inclusion from the very beginning, not as some kind of trickle-down thing that sort of we hope happens well? And I think that means that you ask different questions about who research is for, how it should be funded, where it should go. And that means not only to companies, that means to communities.

That means not only to researchers, it means to different types of research communities, for example, not only elite research communities, I would say. So it is potentially an opportunity, despite, you know, in the midst of all the chaos and devastation and arguably illegality, to, I think, imagine a future for science and technology policy that is more civic, is more community centered, is more publicly engaged and more publicly accountable. but we have to be able to let loose, let go of another man's powerful vision of the future, Vannevar Bushes, right? So that was also future, that was a future making, right, as well. So we're at time, but I'm just going to choose a last question that is kind of an amalgamation of a few different questions in the chat, which is a request, I guess, to end on an optimistic note. And the way I want to put it to you is so, you know, thinking about, I guess we're anti-private manifestos, but your own vision of the future, the future that you desire for those you love and their children and families and the people that are coming after you.

Where do you see hope for that future now? Like where are the places in our multidimensional landscape where you see a possibility for that future? You know, thinking about technology in particular and science and technology policy, I think the fact that in a moment of contestation and polarization, at a moment where, you know, we're being told that in the United States that communities can't agree on anything or people can't agree on anything, that, you know, there's been agreement that the future that is on offer from Silicon Valley, you know, through and with the use of technology is not one that people want. And so the ability to, I think, across a lot of different differences to articulate that that is not what we want. And implicitly, and hopefully that's something that can be, I think, heightened and made more explicit, is if this is the negation, we don't want this, that means we want this. That means that we want technology companies that are accountable to the public, that we want to choose whether or not these technologies are used in our young people's classrooms, that we want to in the same way that we choose screen time at home or not, for example, that, you know, we want to have some say in the pace of how all of this change is happening, that we don't want technological change, even if it's beneficial to come at the cost of our natural resources and of rising pollution.

So I think those are, you know, what gives me hope is that there is a common sense, you know, dare I say it, kind of tension table sense that these are things that people want. So they might not want them in the same way. They might quibble about how you reach those ends. But those are those kind of aspirations. And I think people have rightly been critical of things like the Facebook Oversight Board, you know, its own, you know, Supreme Court and things like the cloud constitution that take up these sort of democratic vehicles, you know, as a, you know, in the sort of private sector role of technology. But I think what they and what something like the AI Bill of Rights, you know, four years ago and to this moment sort of signal is that we know that we need to reach for a language that is, you know, above the individual, that is kind of at a level and also above the technology. A lot of our policy conversation around technology is about the literal tool. Like, so, you know, now we have agents.

What do we do about that? What do we do about the large language models as opposed to what's the true north? Like, what's the world that we want to live in? And then how do we reverse engineer from that to forms of governance and rules and laws and policies and the cluster of organizations that form society from firms to schools to homes? We had so many fantastic questions from the audience, more than we anticipated, and certainly of great quality and helped us to open this conversation up in a great way. So I'm grateful to all of you who joined. Grateful to our colleagues, Ashley Baylor at Tech Policy Press and Mariana Poyares at the Privacy Center for all of your help in putting this together. And, of course, to my colleague at the Privacy Center, Emily Tucker, thank you so much. and to Dr. Alondra Nelson, thank you so much for kicking this off with us. We're very grateful to you for doing this. I was so pleased to be invited and honored and glad to be here. And thank you for dealing with my slight head cold.

I hope it didn't muddle things too much. Absolutely. We expect to do these about every two months. So we're going to try to do a bi-monthly cadence on these when we can attract folks who can challenge us at the level that Dr. Nelson's done today. So grateful to all of you for joining and look forward to the next one. That's it for this episode. I hope you send your feedback. You can write to me at justin.techpolicy.press. Thanks to the Privacy Center. Thanks to my co-founder, Brian Jones. And thank you for listening. Tech Policy Press