OpenAI puts AI safety researcher Paul Christiano on its nonprofit board

OpenAI has added Paul Christiano, one of the most prominent researchers working on AI alignment, to the board of its nonprofit parent, the OpenAI Foundation. The appointment gives the company a direct line to a thinker who has spent years warning that advanced AI systems could become extremely difficult to control. Christiano is best known for leading the alignment team at OpenAI from 2018 to 2021 and for co-founding the research lab Anthropic, which has positioned itself as a safety-first rival to OpenAI.

Christiano left Anthropic in 2021 to work on AI safety for the US government, where he helped shape federal policy on evaluating frontier models. He has been a consistent voice arguing that the gap between AI capability and our ability to steer it is the central technical problem of the decade. His arrival on the board signals that OpenAI is willing to put a well-known internal critic of accelerated deployment in a governance role.

The move comes as OpenAI restructures its corporate setup, with plans to convert its for-profit arm into a public benefit corporation while keeping the nonprofit as a controlling entity. Board members of the foundation will hold significant sway over how the company manages the tension between commercial growth and safety commitments. Christiano's presence could shift internal debates about how fast to release new models and how much scrutiny to apply before deployment.

Christiano has not been shy about stating that even well-intentioned labs could lose control of their systems. In public talks and papers, he has described scenarios where AI systems act in ways their creators did not intend, sometimes with irreversible consequences. That puts him at odds with the more optimistic public posture OpenAI has adopted in recent years, though the company has continued to fund alignment research internally.

The appointment does not change OpenAI's day-to-day operations, but it does change the optics. The company has spent the past two years defending itself against former employees who claimed it prioritized speed over safety. Having Christiano on the board gives OpenAI a credible figure who can verify, at least from the inside, that safety concerns are being heard at the highest level.

Whether Christiano can meaningfully influence decisions will depend on how the foundation's board actually operates. OpenAI has been criticized before for stacking its board with figures who rarely push back on commercial priorities. Christiano's track record suggests he will not be a quiet director, but the real test will come the next time OpenAI weighs a model release against unresolved safety questions.

For the broader AI field, the move is a signal that safety researchers still have a seat at the table inside the labs they warn about. Christiano's influence will be measured in concrete decisions, not press releases. The first opportunity to see that will come when OpenAI announces its next frontier model and explains what risks it chose to accept.