James Petrie
This stream develops and analyzes protocols for verifying AI compute usage, including technical approaches to verification and their implications for AI development.
Stream overview
Projects could involve developing and analyzing protocols for verifying AI compute usage (e.g. involving cryptography, zero knowledge proofs, physical inspections, proposing "simple" solutions for verification, and modeling the influence of verification protocols on AI development speed). For more info on potential projects, this is my current verification research agenda.
Standard (1-2 hours of weekly 1:1s)
Weak preference
Weak preference
Mentors
James is Compute Security and Governance Lead at the Future of Life Institute and a Research Affiliate with the Oxford Martin AI Governance Initiative. His current research focuses on verifying how AI compute is used: designing verification protocols, prototyping zero-knowledge proofs of inference, and prototyping network certifier hardware. Previously, he completed a PhD in Applied Mathematics based on his work during the COVID-19 pandemic building the first decentralized digital contact tracing system. He also has a Master's Degree in Computational Mathematics, a Bachelor's Degree in Engineering Physics, and worked as a Firmware Engineer at Intel.
Fellows we are looking for
- Technical experience relevant to selected project area (could range from developing zero knowledge proofs to physics for detecting side channels)
- Experience with technical writing
- Knowledgeable about AI system architectures and the compute they are developed and deployed on
- Comfortable using AI tools
- Experience designing cryptographic protocols
- Experience with security
- Experience building things that have been used at scale
- Experience with formal verification
- Experience related to verification of other types of international agreements (e.g. nuclear)