OpenAI Puts Paul Christiano on Its Foundation Board

OpenAI named Paul Christiano to the OpenAI Foundation Board on 9 September and placed him on its Safety and Security Committee, the body that, in OpenAI's words, "provides governance over safety and security practices across all of OpenAI, including OpenAI Group PBC." He will also sit as a non-voting observer on the PBC board itself.

Christiano joins the committee alongside its chair, Zico Kolter. The SSC is a committee of the Foundation Board rather than of the company, and the Foundation both controls OpenAI Group PBC and holds a significant equity stake in it — the structural detail that separates this seat from an advisory one.

He is an unusual appointment by the standards of corporate boards. Christiano led alignment research at OpenAI from 2017 to 2021 and did foundational work on reinforcement learning from human feedback, the technique that made ChatGPT usable. He left to found the Alignment Research Center, a nonprofit focused on aligning advanced AI with human interests, and is currently a senior technical advisor at the Center for AI Standards and Innovation inside NIST, the Commerce Department agency that evaluates frontier models — including for capabilities with national security implications.

He is also, on the record, alarmed. Christiano has spent years arguing that rapid capability gains could produce a catastrophic and irreversible loss of control, and that developers must prepare long before systems reach genuinely dangerous levels. OpenAI did not sand that down in the announcement. It wrote that he "has been an independent voice on whether the industry's safeguards are adequate," and that the Foundation's governance "is stronger with people prepared to challenge prevailing assumptions."

Board chair Bret Taylor framed the hire around technical judgment rather than reassurance, saying Christiano's government work on frontier AI safety and standards "will strengthen the Board's oversight and the Safety and Security Committee's work as OpenAI continues to develop and deploy frontier AI."

The dual role comes with a wall built around it. A footnote to the announcement states that in his CAISI capacity Christiano "will recuse himself from all OpenAI-related matters as well as all model evaluations" — an acknowledgement that a government evaluator sitting on the board of an evaluated company is a conflict that has to be handled in writing rather than assumed away. Read one way, the recusal is the minimum the arrangement required; read another, it is one more channel between frontier labs and the people who write the standards.

The appointment builds on the governance structure established in OpenAI's October 2025 recapitalization, including commitments made in connection with the year-long reviews by the California and Delaware attorneys general. Those reviews are why the Foundation ended up controlling the company at all, and the Safety and Security Committee is one of the instruments they produced.

The timing is hard to miss. It lands in the same week that an Anthropic alignment lead put the odds of AI killing everyone above 10% and a researcher there resigned over self-improving systems, and a few weeks after OpenAI redirected a fifth of its compute to safety monitoring. Putting the researcher most identified with taking catastrophic risk seriously onto the committee that governs its safety practices is a bet that those arguments are better made inside the room than outside it.