OpenAI researcher Dan Selsam publishes personal statement on AI risk

thlarsen · x · 2026-09-16

Amid ongoing discussion about concentration within the AI safety community, Jacques Thibs laments that early advocates spent years pushing AI risk awareness only to be dismissed. In the quoted post, Daniel Kokotajlo shares a public personal statement from Dan Selsam, a current OpenAI capabilities researcher (his former boss). Selsam has worked on AI for over fifteen years: probabilistic programming at MIT, early development of the Lean Theorem Prover at Microsoft Research, some of the first demonstrations of neural networks learning to reason during his Stanford PhD, and nearly five years at OpenAI pioneering chain-of-thought optimization. The statement lays out his views on AI risk.

Related event: OpenAI researcher Dan Selsam warns alignment evaluations are failing(24 posts)→

Original post →

More from AGI Musings

AGI Musings channel →