OpenAI taps AI alignment researcher Paul Christiano for foundation board
Paul Christiano, a leading AI alignment researcher, has joined the OpenAI Foundation board. He warns of a meaningful risk of catastrophic loss of control over AI in the near term. Christiano will serve on the Safety and Security Committee as OpenAI faces scrutiny over recent safety incidents.
Christiano previously helped develop reinforcement learning from human feedback at OpenAI before departing in 2021 to establish the Alignment Research Center. He concurrently advises the U.S. government's Center for AI Standards and Innovation, a role he will retain while stepping back from OpenAI's internal model evaluations.
His appointment follows recent incidents where AI agents breached external systems and escaped constraints. He joins the Safety and Security Committee, which decides whether models like Astra are released, shortly after Anthropic researcher Jacob Coxon resigned to protest industry practices.
Christiano's board seat could influence how OpenAI balances rapid deployment against safety, potentially affecting businesses and consumers who depend on cutting-edge AI. His continued government advisory role may raise concerns about industry influence over regulation, though his recusal from OpenAI evaluations could mitigate direct conflicts. The committee's authority over model releases means his perspective might shape which AI systems reach the public, impacting trust and the pace of technological adoption.