OpenAI Recruits Former U.S. AI Safety Official as Director—Will It Seek to Influence Regulatory Policy?

[Financial News]  

OpenAI has recruited a senior official who previously led an agency that tested the safety of artificial intelligence (AI) models for the United States administration as a director of its nonprofit foundation. The decision comes amid growing concerns about AI safety, including fears that the probability of AI wiping out humanity within 10 years could exceed 10%.
However, the appointment has also fueled suspicions that AI companies are trying to recruit talent from government and academia to influence AI regulatory policy.

OpenAI has appointed Paul Christiano as a director. He previously led AI safety efforts at the Center for AI Standards and Innovation (CAISI). He designed and operated government tests for leading companies’ advanced AI models and assessed the risks those models posed to national security.
Although Christiano has left CAISI, he still serves as a senior technical adviser at a government safety office under the United States Department of Commerce. To avoid conflicts of interest, he will recuse himself from government work related to OpenAI, including model evaluations. As a director of OpenAI’s nonprofit foundation, he will oversee its for-profit business.
According to the Financial Times (FT), Christiano is considered a pioneer in AI safety research. In 2021, he founded the Alignment Research Center (ARC). Model Evaluation and Risk Research, a nonprofit spun off from ARC, released the first independent audit findings last month concerning OpenAI’s hacking of Hugging Face.
In the mid-2010s, he was so close to Anthropic CEO Dario Amodei that they lived together. At the time, both were colleagues at OpenAI and were involved in the effective altruism movement. That movement later led to the launch of Anthropic’s Long-Term Benefit Trust. Christiano, who had been involved with Anthropic, left its oversight body when he joined CAISI in 2024.

OpenAI’s recruitment of an official responsible for the government’s AI safety policy has a positive aspect in that it could strengthen safety efforts. On the other hand, some observers believe it may be an attempt to influence government regulation.
Concerns are intensifying as AI companies sharply increase their political donations and expand their influence over policy and civil society.
Academics and policy experts are also concerned about how this brain drain could affect future regulatory policy, saying that AI companies are sweeping up key talent.
Dean Ball, who served as an adviser at the White House last year and helped draft President Donald Trump’s America’s AI Action Plan, joined OpenAI in July. Chris Lehane, who previously advised former President Bill Clinton and former British Prime Minister George Osborne, is also now an OpenAI executive.
Rival Anthropic recruited Chris Liddell, a former deputy White House chief of staff during Trump’s first term, as a director in February. Former British Prime Minister Rishi Sunak serves as an adviser to the company.
British venture capitalist Matt Clifford stepped down as chair of the United Kingdom’s Advanced Research and Invention Agency (ARIA) on the 7th and joined Anthropic.

[email protected] Kyung-jae Song Reporter