Skip to content
NewsLeadership

Paul Christiano joins OpenAI Foundation Board and Safety and Security Committee

· by Pondero Newsdesk

The short version

Alignment researcher and RLHF co-developer Paul Christiano joined the OpenAI Foundation Board and its Safety and Security Committee on September 9, 2026, gaining a formal vote on whether new OpenAI models clear for release.

Paul Christiano joins OpenAI Foundation Board and Safety and Security Committee

Paul Christiano, who co-developed reinforcement learning from human feedback and founded the Alignment Research Center after leaving OpenAI in 2021, rejoined the organization on September 9, 2026 in a governance capacity, taking a seat on the committee that holds final authority over whether OpenAI's frontier models ship.

What happened

OpenAI announced that Christiano joined the OpenAI Foundation Board and its Safety and Security Committee. He will also serve as a non-voting observer on the OpenAI Group PBC board. The Safety and Security Committee, chaired by Carnegie Mellon professor Zico Kolter, holds final authority over releasing new models, per TechCrunch.

Christiano led alignment research at OpenAI from 2017 to 2021 and contributed foundational work on RLHF, the technique that shaped instruction-tuned language models across the field. He left to found the Alignment Research Center (ARC), a nonprofit focused on evaluating whether AI systems could undermine their creators. His current affiliation with the Center for AI Standards and Innovation (CAISI), a NIST program, means he will recuse from OpenAI matters related to that role.

In a statement tied to the announcement, Christiano said: "I now believe there is a meaningful risk that rapid acceleration in AI capabilities leads to catastrophic and irreversible loss of control in the very near term," per TechCrunch. He added that he does not believe the AI industry, including OpenAI, is currently on track to reduce that risk to an acceptable level.

Why it matters

The Safety and Security Committee's veto over model releases makes this operationally significant for everyone building on OpenAI's API. A researcher who publicly flags catastrophic near-term risk now sits on the body that gates when GPT successors get cleared. That cuts both ways: Christiano's presence could slow capability releases when a model fails safety evals, or it could give the evals process enough credibility that cleared models face less external scrutiny.

The appointment arrives as OpenAI works to complete its transition from a capped-profit structure to a public benefit corporation. Adding a safety-credentialed outside voice to both boards appears calibrated to address sustained criticism that OpenAI's governance subordinates safety decisions to commercial velocity. Christiano's public record of criticizing the pace of the broader industry, including OpenAI, is a deliberate signal. Whether it translates into structural influence depends on how the committee votes when a borderline model evaluation surfaces.

For AI-tool operators whose products rely on OpenAI models, the practical question is timeline predictability. A committee with a strong safety mandate can extend evaluation cycles, which in turn delays API access to new capabilities.

What to watch next

The first model evaluation cycle after Christiano joins will be the clearest signal of whether his role carries real decision weight. Watch also for any shift in how the Safety and Security Committee discloses evaluation outcomes, a step Christiano has advocated for in past writing on AI transparency.

Sources