PC Paul Christiano On Dwarkesh Podcast

“I think my current view is that this dynamic seems bad enough that right now, from a national security perspective, it is not good for U.S. AI labs to be improving pre-training and reinforcement learning given their current security posture.”

Dwarkesh Podcast · AI Research & Frontier Labs · November 2026

“I think my current view is that this dynamic seems bad enough that right now, from a national security perspective, it is not good for U.S. AI labs to be improving pre-training and reinforcement learning given their current security posture.” — Paul Christiano, Dwarkesh Podcast

Context

Christiano had laid out how stolen model weights or algorithms would leave a lab unable to slow down safely. Asked what frontier labs should do, he gave this answer, later stressing he speaks as an individual and not for OpenAI, whose board he sits on.

Transcript

Dwarkesh Podcast
Paul Christiano

…posture of rapidly growing startups is not really at the same level. And at the same time, we are currently in a very challenging moment for cybersecurity, which is very hard for defenders in general, given the number of vulnerabilities that are being discovered across the entire software stack.

Dwarkesh Patel

So what should the frontier allowed to do then?

Paul Christiano

So I think we are in, again, this really becomes a problem if you get to systems, if you have rapidly increasing. Capabilities and you produce AI systems which you cannot adequately control, and you would like to, or you do, slow further development or deployment. If at that point the stakes are very high, the systems are very powerful, you really do not want it to be possible for an adversary to have access to either the models or the algorithms. I think my current view is that this dynamic seems bad enough that right now, from a national security perspective, it is not good for U.S. AI labs to be improving pre-training and reinforcement learning given their current security posture.

Dwarkesh Patel

Pause.

Paul Christiano

I think we should pause can mean a lot of things. So I think there's like a lot of good things that Frontier Labs are doing that they should be doing. But I do think we like improving algorithms in a way that can be very, very easily diffused, even normally, and could be quickly stolen. Yeah, I…

Dwarkesh Patel

you're on the OpenAI's board. I'm not

Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.

More from Dwarkesh Podcast