Jide Alaga

Projects

Research Direction: AI Safety Coordination

My preference is for the fellowship to focus on creating a fleshed-out proposal related to AI safety coordination between companies. This doesn’t need to be a formal paper, though I’m happy for a mentee to turn it into one if they'd like. We would decide on a specific topic together based on what suits them best, but here are some early sketches:

  • Operationalizing a commitment to allocate a higher ratio of resources (e.g. money, compute, headcount) to safety research once specific R&D thresholds are met.

  • Designing a concrete proposal for how a coordinated slowdown of AI R&D automation might work in practice.

  • Proposing what the next generation of voluntary safety commitments should look like and how to keep them effective.

What I'm looking for in a Mentee

I work best with mentees who are very transparent about their blockers and uncertainties. In particular, I have a preference for an autonomous working style where you take the lead on tasks but are proactive about checking in whenever you need something or notice a block. Ideally you'd also have a good 'radar' for what is realistically achievable within the current policy Overton window.

What I'm Like as a Mentor

I prefer informal communication and am much more responsive on Signal than on Slack or email (I can usually get back to you by the end of the day on Signal). We’ll check in twice a week to resolve blockers and discuss your progress, though I’m happy to adapt our call format to whatever works best for you. I consider myself to be very patient and will often point you toward the right resources so you can find information yourself, but I can also be hands-on with research and writing when I have time.

Bio

I’m a Research Scientist on Google DeepMind’s AI Safety and Alignment team, where I focus on AI governance policies relevant for our Frontier Safety Framework. Before this, I was a member of policy staff at METR and spent time as a Research Scholar at the Centre for the Governance of AI, where I’m now an affiliate.

Previous
Previous

Peter Hase

Next
Next

Samira Nedungadi & Jasper Götting