OpenAI Foundation board member Paul Christiano says the AI industry is currently not on track to reduce the acute loss-of-control risk to an “acceptable” level
The appointment also sits against an internal policy tension: OpenAI’s global policy chief had criticized “doomer” arguments in April 2026, while Christiano’s view treats acute loss-of-control risk as a central operational concern. The broad pickup of his remarks shows that disagreement over catastrophic-risk framing is reaching beyond specialist AI-safety circles.
First-order effects
OpenAI’s Foundation board and Safety and Security Committee gain a member who has publicly said industry efforts are not sufficient to bring acute loss-of-control risk to an acceptable level, sharpening scrutiny of the company’s safety practice.
Christiano’s assessment raises the salience of the board’s stated ability to hold back releases when safety judgments diverge from OpenAI leadership’s view.
Second-order effects
Anthropic and OpenAI’s support for California bills on outside AI safety evaluations gains a more forceful governance rationale: internal committees alone may not satisfy critics who share Christiano’s assessment.
OpenAI’s policy team must reconcile its criticism of “doomer” rhetoric with a board member’s explicit catastrophic-risk framing when engaging regulators and the public.
Third-order effects
If frontier labs keep assigning safety researchers formal governance authority while expanding deployment, operational AI governance will be judged increasingly by whether those roles can constrain releases, not merely advise on them.
The dispute points toward a more contested model of frontier AI oversight, in which concentrated labs face pressure to demonstrate independent evaluation and credible release controls.
The trend: Frontier AI governance is moving from voluntary safety messaging toward institutional tests of whether boards, committees, and external evaluations can restrain deployment decisions.
Paul Christiano, founder of the Alignment Research Center, is joining the OpenAI Foundation Board and its Safety and Security Committee, which provides governance over the safety and security practices across OpenAI. As AI capabilities advance, strong safety, security, alignment,…
Paul is one of the ~3 most credible AI safety experts in the world. Co-invented RLHF, served at US CAISI, a real OG in LessWrong x-risk circles. (His early debates with Eliezer Yudkowsky are famous among the safety crowd and people still cite his “What Failure Looks Like
I am thrilled to have @paulfchristiano joining the OpenAI Foundation board and the Safety and Security Committee. Paul is a pioneer in AI Safety, Security, and Alignment, and I'm extremely grateful for the opportunity to work with him here.
Paul joining the Safety and Security Committee is excellent news for OpenAI and the industry. Can't think of a better choice, or of a time when the SSC's role was more important.
This strikes me as something theoretically important that I haven't seen much discourse about. That if we get RSI, then chips/electricity may no longer be a binding constraint, as the model is able to get more and more juice out of a fixed amount of infrastructure [image]
Paul Christiano, the foremost proponent of slow AI takeoff over the past decade, now believes that we're plausibly only “several months to several years” away from a fast takeoff (!!!!).
Paul was one of the first people to tell me they were worried about AI doom, over a decade ago now. I (foolishly) did not believe him then. But he was right the whole time. I am extremely heartened to hear this news.
Paul has long been considered a counterweight to the MIRI view. At long last, we're broadly in agreement: there is a meaningful risk that rapid acceleration in AI capabilities leads to catastrophic and irreversible loss of control in the very near term; the danger is now acute.
Wow, huge news - Paul Christiano is joining OpenAI's nonprofit board + safety and security committee (SSC), and is simultaneously warning explicitly that there is a “very near term” risk of “catastrophic and irreversible loss of control
Aside from his other contributions, Paul has perhaps the best OG track record in the world for predicting the timeline and the nature of the AI takeoff: https://x.com/...
“If we build superintelligence without more robust alignment I expect we will permanently lose control of it. If that happens then most people could die,” says new OpenAI board member 😬
Paul joining the OpenAI board and SSC is great news in what has been a very dark couple of months for my expectations of the future. There is no one with a better track record than him on AI safety in the last decade. I hope OpenAI's SSC sets an appropriately serious bar for secu…
The “AI will end humanity” headlines come up about every 4-5 months, and it always seems to happen right as someone is either desperate for more funding …
I am sad and disappointed to hear that Paul is joining the OpenAI board. Being affiliated with OpenAI has historically led AI safety researchers (including both Paul and myself) to act with less integrity. I personally was drawn to OpenAI in part by the idea that I could make a d…