Analysis
OpenAI added Paul Christiano, an influential AI alignment researcher, to the board of the OpenAI Foundation, TechCrunch reported September 9, one day after Pulse covered Anthropic and OpenAI safety researchers publicly estimating double-digit odds that AI causes human extinction within a decade. Christiano will sit on the board's Safety and Security Committee, chaired by Carnegie Mellon professor Zico Kolter, which holds final say on whether OpenAI releases new models.
What's changed since Tuesday's story: rather than just researchers voicing personal risk estimates on social media, OpenAI has now made an institutional governance move in direct response to the same debate. Christiano previously ran OpenAI's own alignment team before leaving to found the Alignment Research Center, and has spent years publishing detailed technical arguments for why AI development poses catastrophic risk -- credentials that made him one of the more credible "doomer" voices to install in an actual decision-making seat rather than just publish papers from outside.
“He said he joined anyway because he believes OpenAI could meaningfully cut the risk if it chooses to.”
In his own statement, Christiano wrote that he believes there is "a meaningful risk that rapid acceleration in AI capabilities leads to catastrophic and irreversible loss of control in the very near term," and said he does not think the AI industry, including OpenAI, is currently on track to reduce that risk to an acceptable level -- language nearly identical to what Anthropic's Evan Hubinger and Samuel Marks said in the same news cycle earlier this week. He said he joined anyway because he believes OpenAI could meaningfully cut the risk if it chooses to.
The appointment reads as OpenAI trying to convert a public-relations problem into a governance credential: adding a credentialed skeptic to a board seat with actual model-release veto power is a stronger response than a blog post, but it's also a low-cost one if Christiano's committee vote can still be outvoted by the rest of the board. Whether this changes any actual release decision, rather than just OpenAI's messaging around the extinction-risk debate, is the open question Christiano's first contested vote will answer.
