Liz Barry: Engineering Democratic Systems for the Age of AI | Intelligence at the Frontier 2026

Um, hi everyone. Let me get settled here. Um, and hi to folks that I know in the crowd and if I haven't met you yet, I look forward to meeting you. Um, I'm thrilled to be here this weekend in this experimental tower of frontiers. Um, my name is Liz, Metagov's executive director.

And I'm delighted to welcome you to the Metagov block. Um, both here and in other floors, you can find Metagovernors such as SJ Klein curating floor 16 and tomorrow Seth Frey can be found on the AI floor. So, Metagov is a non-profit research to infrastructure laboratory with diverse research leadership in self-governance. Um, just quickly, we uh, let's see, how do I click this? Great.

Um, we've analyzed a billion dollars of investment into digital public goods. We created standards that are governing 19 billion in on-chain value. We launched publicai.co an inference utility with Switzerland and then built a co-op around it. And me personally, I'm involved in co-directing two million dollars of investment into democratic infrastructure and advising on 10 million dollars more.

So, uh, our

[cough]

our mission [clears throat] is to describe, support, and expand the possibility of self-governance in a digital age on the internet and beyond. We advance distributed, accountable, community-driven governance in digital systems by providing the scaffolding necessary for emergent institutional design. We aim to reduce the friction of creating new organizations that are alive, that can adapt, incorporate feedback, and manage power. Metagov's also a learning community dedicated to growing our combined capacities to go out in the world and improve our local governance wherever we may be. Um a little provocation is what determines if technology is net good or net bad?

It's not about its functionality. It's about who controls it. And so to figure out how to share control, we have to figure out governance. So I'll be sharing stories from the field um and then I'll be handing off to an excellent series of speakers that are going to demonstrate through their work and leadership um capacities uh emerging capacities that are capable of rising to the challenge of governing AI and planetary health locally worldwide. I'm just going to stay on this one image for now.

So this was Sunday afternoon in Los Angeles. Um 3 years ago there was a there was corruption in city council. There was a racist tape recorded among three city council members colluding to keep power even in the face of redistricting meant to reduce corruption. And um that's actually not the only corruption on the Los Angeles city council, but I don't have time to tell you about all of it. Um but this recording forced the city council to begin a process where other people revised their power.

In fact, the document that they were that they had to revise is called the charter of the city of Los Angeles. And there were a dozen or so folks appointed to that commission. But once those people realized what they were up against, they reached out to civil society including Metagov and they Foundations in Los Angeles to help engage the people of LA in their own constitutional process. So, that's how a group of facilitators, democracy activists, journalists, and even the Hollywood director of the film Everything Everywhere All at Once showed up to help out the local government of Los Angeles. Um we hosted online deliberations so that the people could set the agenda for mini assemblies and ultimately the full citizen assembly.

And we gathered the people's vision into a preamble to the charter. To do this, we had to diversify our team, which is a theme I'm going to be repeating again and again. We had to diversity our diversify our team so that we could engineer at the level of complexity required to shore up democracy. We needed a print shop to print invitations to participate. We needed geographers to get us a list of every address in Los Angeles.

We needed the post office to deliver 20,000 postcards. And we needed specific nerd-doms who could come in and teach the delegates what research shows about voting systems and previous attempts to reform um authority. So, the delegates from the population of Los Angeles shown here um met for four full days of briefings, discussions, and deliberations. And on Tuesday, like 3 days ago, we went to City Hall and presented the recommendation that for the first time in 100 years the size of the city council should be expanded from 15 to 25 people. That is something that local democratic processes have not been able to handle for a hundred years.

Um we also said, maybe kind of a no-brainer, that the council people and the mayor shall not be able to appoint their own ethics oversight board. Keep in mind, none of this content came from me. I ran the processes and I'm happy to promote the results, but the people coming up with these answers were from the population of Los Angeles. Now, I work overtime on democratic upgrades like this all around the world. For me, 2025 and 2026, it's like being deployed in the fight against democratic backsliding.

I build teams that build tools and infrastructure for large-scale democratic deliberation. I work across community organizing, open-source software, and governance theory. And on a tech side, identity is a frontier for large-scale deliberation systems. Um currently, as many articulate people have already said this morning, deliberative platforms need proof of humanity to show that their results are deliberate. But participants need data sovereignty for safety, privacy, and credible exit.

These two needs are interfering with the integrity of systems that could be the next generation of democratic systems. So, it's really great to be here with you in person to connect some of the philanthropic efforts I'm working on um with some of your initiatives here. Um how I started years ago was with a simple sign that said talk to me and then organizing parties for a thousand people at a time to self-organize themselves around questions. That led to a series of risks and adventures that I'll save for another time, but it brought me to being physically, personally present in Taipei for Taiwan's Sunflower Movement in 2014. And I became part of the hundreds of thousands of people protesting decisions that their elected officials made against Taiwan's own interest in favor of China.

There's a story here about breaking the story to the international community, setting up an organizations with the creator of Polis, working globally to support peace builders and governments, um social movements. Anyways, again, more stories for another time. Um what's going on now is that I've joined Metagov, um where Aviv Ovadya has created a visionary program in interoperability in this kind of deliberative tooling. Um he made an early and decisive move to prevent capture and make sure that emerging AI sense-making capabilities belong to everyone. There are four more speakers after me, and you're going to be hearing about these emerging capabilities, but I'm going to give you a bit of a global overview of what's happening where.

So, even as democratic backsliding is happening, it's also clear that any group's interest in strengthening their capacity to govern themselves is greatly clarified when someone else says that they own you. So, China is saying this to Taiwan. Russia is saying this to Ukraine. The UK is saying this to Scotland. And the United States is saying this to California.

Um and all of these reason regions that I've just mentioned have governments that are driving innovation into collective self-governance capacity. So, in Taiwan, they're developing capacity for democratic steering under duress. They've set up a government-funded institute where staff can get market salaries. Um, they have built participant wallets to protect their citizens' identity even as they enter into deliberations where they're sharing vulnerable political speech. And overall, they're building civic infrastructure commensurate with the risks.

Um, in Kenya, where their agency protests still roiling for 3 years and counting against President Ruto, um, youth have engaged peacebuilders in radio- and text-based engagement to clarify the youth demands, m- help mature their movement, and bring about a peaceful civil society. In Scotland, they're working on developing their self-governance capacity while being inside of an empire that controls most of their land. What Scotland did was create an accelerator program to build the company they wanted to procure open-source software from, and then procure it, so that they could run national conversations on AI regulation, societal health, and maybe eventually independence. Um, I have stories from the UK, Slovakia, Finland, and right here in California, the Engage California team, which had started working on youth mental health, but then had to divert to responding to the wildfires of just 1 year ago. And how they did it was they had to pull the skilled staff needed from across departments that didn't otherwise work together.

Um, they're still trying to California is trying to figure out how to institutionalize collective intelligence processes. And so, I just want to return to the theme of like, how do we build the diverse teams needed um, that can support um, democratic systems that fit for purpose. So, a century ago, Hannah Arendt called it. She said deliberative democracy was not going to be able to scale to mass society. But, taken from another point of view, democracy's time has not yet come.

Our attempts at of, by, and for have been starved for the capabilities that we are just developing now. Democratic institutions are most commonly torn down by the from the inside by people frustrated with a combination of corruption and the inability to handle complexity. That should be an aha for the people in this room. We know how to bring transparency to data, and we know how to apply tools to handle high-dimensional situations. And yet, we can't do it alone.

We as technologists don't possess the epistemic diversity required to build robust multi-stakeholder systems that any fraction of our species would actually use. So, along with Metagov research director Seth Frey, I'm obsessed with governance as a thing that people do to develop themselves, and enter a mode of associated living, and it matters which part of these systems we automate. There's a paradox of automation that if we reach for tooling to augment a fundamental weakness, when that tooling is later removed, the weakness remains and may have further weakened. So, digital democratic scaffolding, such as we're welcoming millions of people into even in this in 2026, should leave people stronger as they head back into their lives, more capable of engaging with complexity and change. And my time is now up.

Um I I pose one last question. It If my first question was who is controlling the technology, my last question for today is one that connects to the vision of this tower. What was said earlier about moving from community to society? What if we let the same randomness that brings rigor to clinical trials, cryptography, statistical research, and risk modeling into building teams and movements capable of making change? What if the protocols we need for human flourishing are actually very simple?

The protocols for the way we associate with people who are not like us, so that together we can build systems fit for democratic purpose. Thank you.

[music]