OpenAI wants outside auditors

OpenAI supports parts of a bipartisan bill that would make the largest AI companies submit their models to independent audits. It is the first federal mandate of this kind the company has backed.

Pitched straight to lawmakers

Chris Lehane, OpenAI's chief global affairs officer, voiced support at a roundtable with lawmakers in Washington on Tuesday. The proposal sits inside a bill called the FRONTIER Act.

What auditors would judge

An independent group would have to confirm that a company's safety protocols are strong enough to keep catastrophic risk within limits the bill sets out.

“acceptable levels of catastrophic risk”

More than audits in the bill

Sponsored by Reps. Jay Obernolte (R-Calif.) and Lori Trahan (D-Mass.), the bill would also make developers publish reports on their models and disclose safety incidents. OpenAI has not endorsed all of it.

Why OpenAI wants rules now

In a Sept. 9 blog post, Lehane urged Congress to set mandatory national AI safety requirements and called on leading labs to build shared industry standards together.

“the capabilities that make models more useful also come with risks, and they will not remain confined to a few frontier laboratories”

Safety advocates want government in charge

Joseph Miller of PauseAI UK says the debate is long overdue. He doubts AI labs will take strong enough safety steps on their own without the government stepping in.

“We need government to lead on it”