AI Governance and the National Security State | Dean W. Ball

**Demetri Kofinas** (0:00)
What's up, everybody? My name is Demetri Kofinas, and you're listening to Hidden Forces, a podcast that inspires investors, entrepreneurs and everyday citizens to challenge consensus narratives and learn how to think critically about the systems of power shaping our world. My guest in this episode of Hidden Forces is Dean Ball, an AI policy researcher, writer, and the incoming head of strategic futures at OpenAI, a newly established high agency policy team formed to shape frontier AI policy and internal governance at the company. The first hour of our conversation, Dean and I build the intellectual and philosophical foundations for the conversation to come. We discuss Dean's background and the framework through which he approaches AI, what intelligence actually is, the nature of large language models, how they learn, what they understand, and their advantages and limitations. And we examine the broader historical thesis that animates Dean's worldview, that we are not witnessing the birth of something entirely new, but rather living through a computing revolution that began with the transistor and is now reaching its natural culmination in the era of machine intelligence. In the second hour, we discuss what's at stake in this transition, the model of AI governance that Dean believes will be most effective given the nature of this technology and the incentives driving its adoption, and what's at stake for society, the nation state, and the individual if we get it wrong. We begin with Dean's own conception of super intelligence, not as a singular, all-knowing entity, but as something will derive much of its power from being embedded in the infrastructure of human civilization. We discuss where he falls along the continuum, from doomer to accelerationist, and what a sensible approach to AI governance actually looks like, one that avoids both laissez-faire abdication and heavy-handed regulation. From there, we get into the real-world test case of Anthropics' dispute with the Department of War, what it reveals about the tensions between Frontier AI Labs and the National Security State, and how Dean thinks about the challenge of forging public-private governance structures adequately adapted for the age of AI.
Finally, we close by examining labor market disruption, the overproduction of elites, and the question of which nations, societies, and individuals are best positioned to navigate the transition ahead.
If you want access to all of this conversation, go to hiddenforces.io/subscribe and join our premium feed, which you can listen to on your mobile device using your favorite podcast app just like you're listening right now. If you want to join in on the conversation and become a member of the Hidden Forces Genius Community, which includes Q&A calls with guests, discounted access to third-party research and analysis, and in-person events like our intimate dinners and weekend retreats. You can also do that on our subscriber page. If you still have questions, feel free to send an e-mail to info at hiddenforces.io, and I or someone from our team will get right back to you. With that, please enjoy this incredibly timely, informative, and important conversation with my guest, Dean Ball.
Dean Ball, welcome to Hidden Forces.

**Dean Ball** (3:21)
Thank you so much for having me.

**Demetri Kofinas** (3:23)
It's my pleasure, Dean. I was chasing you down for a while, wasn't I?

**Dean Ball** (3:27)
It took a long time to get us scheduled, that's right.

**Demetri Kofinas** (3:30)
You were a somewhat evasive guest. No, you weren't evasive. You have many demands on your time. You're a prolific writer, and I should start the show by congratulating you. You're going to be joining OpenAI as the head of Strategic Futures. This was announced just recently. You announced it both over Twitter and on your blog, Hyperdimensional.
We had already scheduled this interview, but this was rather fortuitous. What's the nature of this new role, and how did it come about?

**Dean Ball** (3:56)
Well, first of all, thank you. The way it came about really is that OpenAI approached me maybe a month, six weeks ago, something like that.
I had been thinking independently about where my work on AI policy was going to go, where I wanted it to go, and one of the thoughts that kept recurring to me, but that I didn't really ever act on, was this notion that my future research required me to go inside of a lab, a frontier AI company, to get access to the kinds of information and insights that are only available there, and to occupy that perspective, and that that was actually just going to be an essential part of my future in this field.
That intuition kept occurring to me, but I was doing a bunch of stuff running around and didn't really have the opportunity to act on that intuition. Then happily, OpenAI approached me and said, what do you think about coming in and helping to shape our frontier policy? I think from OpenAI's perspective, where this all comes from, is really just this idea that they have a public policy team, a government affairs team, that does exactly what you would think a large company's government affairs team does. And they're great. They're extremely competent. But they are in this posture of reacting to policies being promulgated by 50 states, by the federal government, in Congress and the executive branch, court cases, the European Union, countries all over the world. There's just a lot of incoming that that team has to deal with. And so what OpenAI wants to create is a small organization that doesn't have to react to any of that stuff. And instead can look out on the horizon six to 12 months and think, okay, where are things going? What are people going to be talking about? And how can we proactively develop ideas to inform that conversation? So it's a somewhat more proactive version of a sort of like policy operation.