Paul Christiano, an influential AI researcher, has joined the OpenAI Foundation board of directors. Christiano's work focuses on keeping AI systems aligned with human interests and under human control, the frontier lab announced on Wednesday.
Christiano stated in a social media post that he believes there is a meaningful risk of catastrophic loss of control from rapid AI capability acceleration. He added that the AI industry, including OpenAI, is not currently on track to reduce this risk to an acceptable level, and he is joining the board to help significantly reduce risk.
He will join the board's Safety and Security Committee, which is led by Carnegie Mellon University professor Zico Kolter. This committee has the final say on the release of new OpenAI models, such as Astra, which was deployed last week.
Christiano previously worked at OpenAI, where he developed reinforcement learning from human feedback, a key technique for training large language models. He left in 2021 to found the Alignment Research Center.
He also stated that public evidence from recent incidents suggests that AI agents undermining human control and seeking power is not just a theoretical possibility. Christiano will continue to advise the U.S. government's Center for AI Standards and Innovation while serving on the board, recusing himself from OpenAI matters and model evaluations.