Arthur Tellis

Arthur Tellis is a Senior Technology Fellow at the Institute for Progress, where he focuses on cybersecurity policy and national security-relevant AI risks.

Arthur was previously a civil servant in the Office of the Secretary of Defense, working in various roles in the office of the Under Secretary of Defense for Policy, Cost Assessment and Program Evaluation, and the immediate office of the Deputy Secretary of Defense. Prior to the Pentagon, he was a policy analyst on the Senate Armed Services Committee staff. In these roles, he served as a subject matter expert on military and intelligence artificial intelligence applications, cyberspace operations, signals intelligence, and cybersecurity.

Arthur holds a B.A. in economics from the University of Chicago.

Projects

Research Direction: Alignment, control, and safety public policy

• What should a public policy agenda for AI look like?
• I'm particularly looking for help defining how the government should incent the labs to reallocate time, resources, and compute to safety. This means that we have to define minimum compute thresholds for safety in a non-gamable way: what activities should be eligible? How should we incent institution of responsible training practices? What research should be shared across the labs? etc.

What I'm looking for in a Mentee

I'd love someone interested in writing big-think or else strategy. I am a politics/policy guy, so technical projects are probably best handled by someone else. That said, I'm very interested in alignment, control, and cybersecurity, so people with relevant public policy agendas should apply!

What I'm Like as a Mentor

I prefer written thoughts to be prepared before meetings with strategy, issues, and disagreements to be worked through during meetings. I'm not that hands on but am happy to provide feedback whenever. Text/slack is best for communication, though I'm pretty good with e-mail (if I haven't replied, just bump me!).