Kyle Redman

The AI & Democracy Foundation is hiring for Program Director / Project Leads, AI Strategy Lead / Director, and Chief of Staff! Learn more and apply: airtable.com/appHnaEzJEzXRd…

AI policy is heading for mainstream politics, and advocates on both sides seem hopeful the tides will break their way. But I'm not so sure: greater salience mostly makes for lower precision and higher volatility. Today, I argue AI politics will be about politics, not about AI.

@Miles_Brundage The popular book I usually recommend is David van Reybrouck's 'Against Elections' This is also something that I, @metaviv and others at @democraticAI are working on. You might be interested in our new 'gap map' for how to improve this approach: democracybuild.org

I've gotten much more bullish on lottocracy / sortition in the past few years. It should be experimented with much more often at least Guerrero's book in particular is pretty well-argued (though too long, there needs to be a popular version of the case for this stuff)

🧵 We are likely to need better democratic systems to navigate the challenges posed by AI advances. Today we're releasing the Democratic Capabilities Gap Map (democracybuild.org) to accelerate work toward making better democracy a reality. democracybuild.org

a small update on what we've been up to with @meaningaligned: we're convening researchers to think about what we're calling "AGI Institutions". we believe roughly that: (1) good futures with powerful AI systems will require an entirely new institutional landscape, a shift of a similar magnitude to the founding of the US; (2) this will require a process of deliberate invention, drawing folk from social choice theory, mechanism design, AI, philosophy, etc; (3) things are moving pretty fast so we better, ya know, get the ball rolling. we hosted a one-day workshop on this in Paris (co-located with IASEAI), and it was a huge success!! we wrote up a little bit about it in a blog post (link below). we'll be running more of these, and also have some other cool projects that we'll talk more about soon :). if you’re working on institution design for a world with powerful AI and want to be involved, reach out to us at research@meaningalignment.org

What should "middle powers" do to secure a place in the AI future? Depends on the country. For Australia, I argue that becoming a data center superpower is the best way forward. 1/8

Open community meetings ≠ deliberative democracy. Only a matter of time before the abundance crowd discovers the role sortition + well-facilitated deliberation can play in mediating positive-sum local outcomes through the likes of citizens' assemblies.

@pawtrammell Conditional on making no improvements to our existing institutions for making democratic decisions. Imagining/developing/creating AI-ready governance/legitimacy-building processes is considerably underfunded compared to all the other failure modes.

Though Fin and I didn’t talk about this—the pod was recorded a few weeks ago—the Anthropic-Pentagon mess seems to me like an example of how it sometimes matters who owns the capital. Seizing Anthropic, or drastically redistributing its ownership, would be different in many ways from taxing its current owners’ consumption. Incidentally, this case also illustrates why, contra what many people who read my essay with Dwarkesh assumed, I personally don’t think it’s at all obvious that more wealth equality would be better. I wish I had flagged this more strongly. E.g. if I had to guess, I’d say that a world with a robot army under the command of Anthropic’s current owners (whether or not they want it) clears the bar of being better than a world with a robot army under the command of the current median American, let alone the current administration. I’m sure someone will take that out of context to suggest I want some horrific revolution, which I don’t, but I’m saying it to emphasize that we have to think about the positive and normative claims about wealth concentration separately.

I think this one needs no further explanation.

@GuiveAssadi @lumpenspace That would be because our existing democratic arrangements are poor. Most people agree our institutions need to be improved (even before their need to keep pace with AI).

I think today’s events are a good illustration of why “democratically accountable AI” is perhaps not what we want

Worth thinking about how to describe what your organization does, in detail, in a series of plain English markdown files.

Check out the paper (to appear in #ICML2025): arxiv.org/abs/2411.09222 And thanks to my incredible co-authors / Democracy Levels Working group @kredman500, @LukeThorburn_, @cquanze, Oliver Smith, @FlynnDevine, @Werdnamai, @SmithaMilli, Manon Revel, @kjfeng_, @amyxzh, @bilva_chandra, @bakkermichiel and @Dr_Atoosa.

You can visit the website: democracylevels.org And go follow my co-authors, who're all doing incredible incredible work. x.com/metaviv/status…

Better evaluation tools will unlock rapid innovation when it comes to processes, technologies and their combinations. That's something I'm working on now, and I would love to hear from anyone thinking about how to measure different kinds of deliberation.

Red alert. We've just published the 'Democracy Levels Framework', which is a paper, framework and suite of tools to help AI labs, policy-makers, researchers, and civil society make more informed decisions when considering delegating decisions to democratic processes.

🎉 ARIA is excited to fund six new Safeguarded AI teams, bringing together experts from across the social sciences, economics, law, and policy, to address the crucial link between advanced AI and society. (1/3)🧵