Paul Christiano joins OpenAI Foundation Board's Safety and Security Committee

OpenAI has appointed Paul Christiano, the AI alignment researcher who founded the Alignment Research Center (ARC) and helped pioneer reinforcement learning from human feedback, to its Foundation Board. Christiano will serve as a non-voting observer on the OpenAI Group PBC Board and will also join the Foundation Board's Safety and Security Committee (SSC).

What's new

• Christiano joins the OpenAI Foundation Board as a non-voting observer on the OpenAI Group PBC Board.
• He also joins the Safety and Security Committee, working alongside SSC chair Zico Kolter. The SSC provides governance over safety and security practices across all of OpenAI, including OpenAI Group PBC.
• OpenAI's announcement notes Christiano currently serves as a Senior Tech Advisor at the Center for AI Standards and Innovation (CAISI), a NIST body within the U.S. Department of Commerce, where he has worked on evaluating frontier models — including capabilities with national security implications — at CAISI and its predecessor, the U.S. AI Safety Institute.
• From 2017 to 2021 Christiano led alignment research at OpenAI itself, where he contributed foundational work on RLHF, the technique now standard for training large language models including OpenAI's own.

Context

Christiano left OpenAI in 2021 to found ARC, a nonprofit focused on evaluating whether advanced AI systems are aligned with human interests, and has since built a reputation as an independent, technically grounded voice willing to question whether frontier labs' safety practices are keeping pace with rising model capability. His appointment comes as OpenAI has been publicizing an unusually dense run of capability claims — a disputed Navier-Stokes proof, a self-reported crossing of a "Critical" cybersecurity capability threshold for GPT-6 Astra, and an internal misalignment-monitoring program — all within the past week.

Why it matters

Because the seat is non-voting, Christiano's appointment doesn't hand him direct control over OpenAI's decisions. What it does is put a known critic of industry safety norms inside the room where the Foundation's Safety and Security Committee reviews OpenAI's own practices, at a moment when the company is under scrutiny for both the pace of its capability claims and, separately, for agent-containment incidents disclosed this month. OpenAI frames the move explicitly around wanting governance "stronger with people prepared to challenge prevailing assumptions" — a signal that the company sees value in visible, credentialed outside scrutiny even if it doesn't come with a vote.