Akbir Khan
Bio
Akbir Khan is a Member of Technical Staff at Anthropic, where he works on the Horizons team focused on building safe superintelligence. He completed his PhD at the UCL DARK Lab under the supervision of Tim Rocktäschel and Edward Grefenstette, with prior academic training in Mathematics and Physics at UCL and Computer Science at Cambridge. His research centers on Scalable Oversight techniques — particularly the use of multi-agent debate to elicit truthfulness from AI systems — as well as AI control protocols and alignment auditing. His work on LLM debate, exploring whether weaker models can assess the correctness of stronger models, received a Best Paper Award at ICML 2024 for the paper "Debating with More Persuasive LLMs Leads to More Truthful Answers." Before his PhD, he co-founded Spherical Defence Labs, an AI-powered API security startup, and also worked as a Research Analyst at Cooperative AI and a Senior Researcher at Tractable.
Links
- Personal Website
- https://akbir.dev/
- Twitter / X
- LessWrong
- akbir-khan
Grants
from Long-Term Future Fund
Discussion
Sign in to join the discussion.
No comments yet. Be the first to share your thoughts.
Details
- Last Updated
- Mar 22, 2026, 1:53 PM UTC
- Created
- Mar 20, 2026, 2:46 AM UTC