Sheila Warren: The Acceleration of Dehumanization | Intelligence at the Frontier 2026

Many of the technologies that we're talking about today literally can be, you know, the reason why they become popularized and mainstream is thanks to your work leading the World Economic Forum creating the center

For the fourth center or what is called the center for industrial revolution? Yeah, it's center for this so but literally you legitimized so many of these concepts that have received industry and not have done a good job of continuing on but before Sheila Warren crypto was seen as a dirty thing. Uh decentralization was back to that. Yeah, was seen as a weird thing and for like, you know, people working in these fringes and also blockchain was seen as something also just dirty and like Bitcoin was not respected but you came in and you just took your entire career's work in building these bridges between software and social wear layers and you set up this entire thing I like the literally this institution at the World Economic Forum that sanctions what technology is real or not and becomes legitimate or not and you know, the industry hasn't done a good job of taking care of that necessarily but I'm curious you know, as you hear about Luke's comment about resource management and we know that like Norway has done a different way of managing their oil. You know, they are not a victim of the intelligence you know, the resource curse but I'm curious how that this makes you think of like what your career.

Now before I let you go, I just have to think like a moment of reflection on how we're managing the commons of everything you've done, you know. Yeah, well I mean look you know, we had a tide for a period of time. We were trying to I always saw my role I think as you know, we talked about this a lot as normalizing new technologies and making them less scary, making them accessible to people, getting folks to understand how they fit into existing streams and also how they were creating new streams and yeah, some of that's been stewarded in ways that I really love and some of that has been steered in in ways I find terrible, you know? But I actually think that the commons questions are really interesting one. And so, Elinor Ostrom, who was cited her earlier, I think David Favor Doll was citing Elinor Ostrom's the value men.

That's actually where I kind of landed some of my discussions I'd done about micro-communities, right? And the idea that

Micro what? micro-communities, right? And so this idea of localized governance that's very hyper local to small communities, and thinking about governance on a blockchain, for example, is not like a I talk about this as imagining scale differently. It's not that you're taking one model of governance or anything else and blowing that up and trying to extend it, you know, everywhere. I think of it as like a language translation issue.

You do something in English, then you just translate it directly into some other language and assume it's going to be culturally relevant, which it never is. You actually have to think about scale in a very small micro way. And so when I think about commons and Elinor Ostrom, part of what she pioneered was this idea that local communities are better. They're simply going to be better at stewarding resources cuz they understand the context that in a hyper local way, as opposed to trying to bring resource management techniques from other, you know, sectors or other, um, countries or other whatever it is that may or may not be relevant in a local way. And I think that's re- You're seeing this, by the way, in a lot of the activism that's happening around data centers that is very specific to communities, right?

To like towards the side where we're just going And driving driving some of those resources actually back to the communities that data centers are not just a terrible thing, but how do we manage data centers as cooperatives? How do we, you know, use like zero off like all the energy? How do All of that. There it is.

Yeah. Well, well, yeah, it's the the entire field of what we call economics, of course, was derived from Elinor's work. And so I think this idea that we have to get back to the basics and really look at grassroots orientations, hyper local communities for everything from stewardship of resources to actually instantiation of new technologies within communities is a really important frame. And what I love about, you know, funding the commons in general and all of the work here is it is rooted in that concept. That something doesn't have to to you know, a decacorn financially or like have billions of users to be a value.

It can be of like extreme value if it's relevant to a smaller group.

So, can you I'm going to move off the stage, but I want to make sure as you move into your keynote that you talk about I think, you know, what a big part of what make those communities so successful when in Elinor Ostrom's, you know, observations is that these communities understood the culture. They understood the dark matter that you can't touch that make communities, you know, successful in the way they manage these local resources. So, I'm hoping that you can, you know, give us like a rallying cry in this moment as we figure out how do we solve for the intelligence curse, but really looking at like can culture play such a play be that actual role. Okay, you could not have teed me up better.

Thank you. Don't decide.

It's because you get like a better front. Uh so, you know, as I was riding BART here, I was kind of laughing cuz I knew we were going to talk about, you know, AI a lot today and well, what can AI do and what can it not do and how is it replacing people and not. And I have to say first of all, I'm deeply curious to see what happens when the freight actually community. A hyper-local community I do not think AI can replicate. So, there is hope.

There is hope in the drunk college students among us um which I, you know, at one point in time was one and well, nostalgia. Um but yeah, I wanted to talk today just about how I see a lot of the challenges that we tend to talk about in silos as culture problems. They're not tech problems, they're broader culture problems. And we could go on and on all day. There's a whole talk I could do about whether tech is downstream of culture or culture is downstream of technology and it all depends on your perspective and what particular vector you're kind of inserting into the conversation.

Um but for me, we have a gigantic culture problem right now that is a global problem and that's the problem of accelerating dehumanization. And to me, we kind of locate that a lot in technology spheres within AI. The idea about workforce displacement or creativity, you know, all these things. Grammarly and the Grammarly lawsuit is a great example of this. But I actually think you're seeing this in real life, in person-to-person interactions.

I have, you know, family and friends in Minneapolis. And so when I think about the acceleration of deportation without due process, that is a dehumanization problem. Fundamentally, that is possible because one group of people that are currently in power have decided that they have the ability to decide who doesn't doesn't belong based on, you know, accent or language or appearance or vibes or whatever it is. And then act upon that. That is a dehumanizing approach and orientation to immigration, right?

And in many, many sectors and I located that in Minnesota for the obvious reasons that everyone in this room is probably familiar with, but all over the world you are seeing dehumanization that is happening in policy sectors, it's happening in conversation that goes everywhere. And so, even in ways that we don't perceive, we are being dehumanized. And so to me, this is again a cultural problem and it is something we have to really be talking about consistently. It's something I think we need to be addressing in every sector that we're really in cuz it can be solved. It can be addressed through governance, through technology changes, through thinking about layers and everything else.

Let me give you other examples, okay? Surveillance tech. I think most of you if you were if you ran into me over the last month, I probably threw, you know, said to you, "Have you seen the Alex Karp, you know, Andrew Sorkin deal book summit discussion?" I happen to be in that room in New York. And it was scary.

It was scarier than, you know, even those of us who know about defense tech, who have friends in defense tech or friends in nat sec. When people are saying everything out loud, which is helpful from the standpoint of, look, we know what we're dealing with, and it's validating from the standpoint of those of us who have been kind of saying these things for a long time. Many who are more expert than I am on these things. But, you know, this is the moment, right? People are not afraid to own that privacy is gone.

So, I I love that we talk about privacy preserving technologies, but there is a fundamental assumption that we had privacy in the technology to begin with, which in some cases was not really true, right? So, we're kind of putting privacy back We're like We're retro-conning, you know, privacy into the narrative, right? Because we know how important it is, and we know that surveillance is eroding it, but we've already had this like slippery slope kind of thing happening for quite some time in technology, but again, also in culture. As we have slowly slowly frog in a boil you know, boiling water, allowed for convenience or other purposes, our data, our our location, our whatever it is to be taken from us, right? In ways that we we you get permission once, you don't really think about it, and I don't mean this room, I mean like the broader society, um you wind up in in kind of scary places.

So, in the agentic field, we talk a lot about agents as a social to somebody's problems, okay? You could have an agent that had a privacy layer where you were basically putting a shield between you and like the data collector provider. Uh we talk about GDPR as being very outdated, you know, regulation, which of course it is, although at the time I would argue it was trying to do a thing. Um but I think that's also too narrow of a frame. Uh and the thing that I have been seeing many of us, I think, not just me, but I'm the one up here, so the thing I've been seeing coming for quite some time is this intersection, right?

So, I'm a crypto person more than anything else still to this day, and working in crypto for a very long time, um trying to explain it to people and what it is and what it isn't. And stablecoins, I always thought, programmable money, right? Digital currencies, uh is a huge unlock. And it's an unlock that can go any different any one of different ways. Already we're seeing stablecoins provide uh remove friction in certain places, or remove friction in ways that allows access into financial systems in the global economy in ways that, you know, we weren't really seeing, right?

So, remittances is a great example, people talk about this, how it was too expensive to get money, you know, home, whatever ever home might be. In some cases, correspondent banks had pulled out of banks of the of the world so you couldn't actually even hop your money through, you know, a million hops to get it into the country you wanted to go to. People were reverting to flying cash, you know, home to their families. Uh now you can do that using stablecoins, using whatever stablecoin you prefer, you can do a direct wallet-to-wallet transfer, and that is massive. It's a huge unlock that we should celebrate.

But programmable money can go different directions, right? So, when you think about the ability to program money, how does that factor in there in the idea that agents can program themselves? So, if you suddenly have a global smart contract, and you have an AI that can basically decide what the outcome and the metrics are that are best for your given parties there, you suddenly have this huge ocean global sea of the transfer of value, right? That is determined by It's deterministic by what agents decide is the right metric or the right outcome. And then we also have this idea, and I think the popular imagination of AI agents is like, "Well, I have an agent."

And nobody thinks beyond that. And like, and even if every single person has an agent, we tend to think, "Oh, okay. Well, that's not that big of a deal because we're going to have like agent-to-agent interactions, and it's just taking me out of the equation. So, I don't have to do all the nonsense I don't want to do. Like, agent will do it for me."

But you're probably going to have an agent that is customized to a variety of different things, right? Everybody will have a bunch of different agents. And then every corporation's going to have multiple agents. And and they're going to segment, right? By customer By their TAM is going to be hyper-segmented, right?

You're going to have like a precise agent dealing with you from that company. So, you have to kind of imagine a world in which it's like this infinity stream of agents that are like at every stage. Like, something is being sliced or diced or extracted or exploited or whatever it is. And that to me is a dystopian world that gets really complicated. And then when you imagine that at every stage of that, I don't mean to be like alarmist, but when you imagine that at every stage of that, the agent has the ability to code for itself a smart contract that then locks you into a pattern of behavior, I don't know that that is necessarily awesome.

Um I don't feel great about it. I feel uh concerned about it, uh to say the least, right? Now, there are moments in history we've had this kind of, you know, explosion of of um opportunity. We tend to call them revolutions, industrial revolution. Trisha was alluding to the C4IR, which was something the World Economic Forum stood up in San Francisco for about, you know, period of I think it was about eight years it existed.

I was there for five of those years. And the idea there was that these technologies would all intersect, okay? And so this concept of intersectional technology is something I pull from critical race theory. It's the idea that you can't look at technology in a silo. So the intersection like I just talked about between AI agents, for example, and a blockchain or a digital currency like a stablecoin or programmable smart contracts, that intersection these are all going to wind up being horizontals.

And we still talk about them and fund them, for the most part, like they're verticals, right? But they're all horizontals. Everyone's going to have an AI layer. Everyone's going to have a programmable money layer. Ever or you know, programmable value layer.

Everyone's going to have these different layers in everything that they do that are interacting. What I think is really great about that is that we also talk a lot in our industry about the privacy layer we're going to put in, right? Differential privacy. What has to be private when and to whom? The thing we don't talk about as much that I started really noodling on that I haven't written about yet, but I'm when I when I get time is differential friction, right?

And so one of the early promises of Bitcoin and of crypto was that we were going to remove all friction. And we talked about that like it was this great thing. And I've been skeptical about that, you know, for a very long time. Um there's no question. I gave the remittances example on purpose because the removal of those frictions in low dollar payments in the remittances context from, you know, a person who is working and sending money home is a good thing in my opinion, okay?

And so that friction, the friction of the bank fee or the transaction fee is a good thing because it enables a better outcome for those individuals who are transacting in this low dollar way. And in some cases, the outcome is merely the ability to even make that transfer in the first place, right? That's good. But not all friction is bad, okay? Certainly anybody who's in net sec would tell you, we like friction.

Friction is great, right? Friction can be frustrating when the user has to like, you know, everyone who's tried to pay anybody on Zelle than like, you know, $5 or what for It's like incredibly irritating that every day you have to go it, you know, it's it's annoying, right? But you can remove and divorce the inconvenience of a user experience without removing the friction that makes a system safer. And so I think we have to talk a lot about what kinds of friction do we want. So to go back to the legal context, okay?

When I was in law school, I I I did some internships work with the Innocence Project. And back when I was in college and high school, I actually worked for CAP, which is the California Appellate Project. And you know, leaving aside political views on the death penalty, I was very I felt that if a state of California is going to execute someone, they darn well better have a lot of process to make sure that is the right person, okay? That is friction. I would say that is very good friction.

It's a dramatic example of friction, but there are less dramatic examples of where friction is really critical to ensuring a system works the way it should. And so something I think we all owe it to ourselves and society to think about and talk about very very consciously and actively and deliberately is friction. Where do we want friction? Because if we are focused on we don't think about it and if we don't think of friction as governance and we don't have think of friction as a layer, right? That we are deliberately embedding into governance, into our tech stack, into our workflows, into our processes, friction is generally going to be gone.

And I really really worry about what will happen in that case. So I'll go back to that example and the dystopian vision that I have in my mind is the frictionless ability of an AI agent that purports to be working for me that is interpreting my desires incorrectly, not through malice, but just through accident or whatever it is or because maybe I secretly don't know what I actually want or what we could make a million arguments about why it's innocuous. Nevertheless, it has the ability to codify in permanent code form something that then determines how I am operating in the world in my digital life. You know, and that is interfacing with someone else's and those small casual errors before you know it, like it's like the butterfly of effects, right? You're like, "Oh my god, how did we get over here?"

And I find that very, very frightening. So, uh in sum, in conclusion, I don't know where I am on time. Um what I would say is I am very convicted. Um I've done a lot of thinking and working in regenerative circles and other places just about what is the key to our collective liberation. Okay, I'm going to be very dramatic about it.

And it's collective action, there's no question about it. Okay, so um I talked about this into my builders, I talk about this as the Voltron, you know, our project dates me a bit. But this idea that everyone's building a widget, put the widgets together, we get something better, right? The the reality of competing with um uh defense tech and surveillance tech uh with Larry Ellison, well, name-check. Uh the idea that we can compete with anyone is going to be able to compete with that is is it is not going to happen.

So, I will refrain from being more uh pointed than that, but collectively we certainly have the opportunity and I think obligation to try to do that. And so, I love events like this where we all come together, we hear what everyone's working on. And what I would like to do and what I'm hoping to do with this talk is just encourage this conversation part of your day. Think about what you're building, how it is removing or adding friction, what that friction might be doing or not doing that could be prosocial in a system, and how you might connect sides to what someone else is building so that together we can ensure that the right amount of checks are put into our system, well, they're being called out or being nice. Thank you.

Thank you, too. So, I have a philosophical question for you. Is that what is your current work now? What are you doing, Ashley? Turning your Ashley productive friction, intentional friction?

Because, like you said, some of that friction is going to be needful, but of course the other, the current, you know, structure that we have is useful. So, where do you see this map? Where do you see that media pointer or what's your word? See that. So, I know a number of words and a lot of what we're talking about in these cases isn't that good guess some of them are already there.

I know what it have which is the start apps. I'm into your companies. I this is a conversation I'm just going to add things to everything. But Trisha she had better be you cuz something I'm most excited about is I just took a picture of the ESGI society she just runs. Which is we're talking about approval control with the control framework.

Actually I love for you to just discuss concept for a moment because I'm really excited about it and I do think it is one of the ways we can address this problem. Do we hear that amazing What is that? It's not That pipes. And that's our Start it up. It's exciting.

Can you give us the chorus of that like okay we're going to jam out to proving control. Well, so I think you know one of the things you were talking about earlier when it comes to letting it leads back to what Luke was talking about is that we have to introduce control pathways to the system. Because right now when you have you have referred to it that if you have agents if you want them to be useful you cannot have them only on device. You know at times they're going to need to leave and they're going to need to go do stuff and leave your own file. Leave your own We're going to have to be in a multi-cloud multi-vendor environment.

And so what's fascinating is that what no one knows is that this technology for offering control is here. Luke alluded to that many of these technologies are in this building and they can get as confusing cuz all of these are called crypto or blockchain of centralization. But what we're calling them right now is to say look there's a whole new category of technology emerging. They're called approval control. And no one really knows about it because it has multiple names.

So, what approval control offers you is something that we all need for this new world. As Jesse spoke at Bridges, you need independence provability. You need cryptographic tamper-proof solutions. So, in the event of IOT 2029, they all have to be both sponsored. And it is if beings do not have this control, you cannot scale your agency infrastructure.

And the five dimensions of that people are some of the issues I talked about at the summit, which is privacy, your controllability, verifiability, and your identity, and your security. And we actually just held about a year or so at our conference where one of our members first-person project, the Linux Foundation, just figured out how to implement one of our members' technology when their solutions on first-person project. It is one of the biggest threats to obey or any of the source project right now is what is a human goes to human, who was an agent, how do you know who that agent represents? Where is that delegation of, you know, control? Where is who authorized it?

And who's the identity you want to hide? And where is the proof of personhood? Or is it proof of relationships? And there are technologies that are called proof of control that are resolving this. And one of them is the first-person project.

And what the Linux Foundation did is that they are implementing first-person project into the Linux kernel. And so, I think that's why it's so important what you're talking about is that the technology is here, but we don't necessarily know about it because we're using different terminology for it. So, I think it starts out hyper-locally in these buildings. And I think I think we're also not meaning that all of this is more competitive than collaborative, right? And that that's just capitalism.

I mean, it is what it is, but, you know, I think that what I loved about the summit on human agency, which Trish is Trish is referencing, which happened 2 weeks ago to last week, um is is again and and things like this is that it really is going to have to be about collaboration. So, the collaboration between first-person project, Advanced AI Society, Linux, um this company called H2H, like it's really about trying to experiment and put all these things together in one stack as one example of a way that this could work. But, what I really love, to go back about proof of control and the reason why I'm so excited to be working with you, Tricia, actually and others on it, um, is it's it's a framework and it will eventually become a set of principles and standards and all of that that is universally applicable, right? So, it's a way of kind of doing that I think of it's a way of kind of doing all of this that we're talking about that does ensure the preservation of whatever or the injection and I would argue preservation in some cases like the ship has sailed, but the injection back to this idea that we need to be verified, right? In in terms of what is happening.

Like, who is doing what, where, when, and how? And if only if you can have that kind of accountability, if only you can have that visibility and transparency, can you create any accountability in a system? If you don't even know, you know, who's doing whatever, it's impossible to actually create accountability, um, and that I think is going to be one of being again a a a big big problem we don't take an approach like this one.

Yeah, I actually I really believe that we have the the tools to figure this out. So, I think if we actually if the world actually knew we have proof of control tech, AI wouldn't be as scary. It wouldn't be as doomsday what it is right now. And so, part of what we need to tell everyone and part of the messaging is that the greatest use case for crypto and blockchain is here and that's securing the AI stack. It's securing the AI you know, agentic orchestration tech stack.

It's what many of you are working on, but we need to be able to have common terminology to be able to define it just like the success of open source. What they did was they took free software and they said it is not anti-capitalist. It's a it's open source and here's how we define it and here's how we make it commercially defensible and here's how we make it legally viable. So, with that, I want to thank Sheila Warren. Um, we're going to move into a 10-minute break and then we're going to come back and move from the morning which is very much about an intellectual framing to really a practical You're going to hear from people who are implementing many of these technologies here.

We have a speaker who flew in from Argentina. So, you're going to hear Well, as a person who is like maybe in a public, you know, office position or if you're working with a government, and if you are trying to build out the social where level, how do you start to integrate this? So, definitely you have to come back from break in 7 minutes. See you soon.