Companies push to establish AI safety standards body; US competition regulator calls move 'highly suspicious'

Legislation faces long odds with midterms approaching and Congress focused elsewhere

Logos of OpenAI and Anthropic. [Reuters]
Logos of OpenAI and Anthropic. [Reuters]

OpenAI has been collaborating with rivals Anthropic and Google on AI safety, even as calls grow across the industry to slow the pace of AI development.

Chris Lehane, OpenAI's chief global affairs officer, said at a briefing in Washington that the company had been working with the two competitors on AI safety for several weeks, Bloomberg reported Tuesday (local time). "It makes sense to work together with safety as the top priority," he said.

Lehane said the three companies do not believe an antitrust exemption is necessary for them to cooperate on AI safety.

Federal Trade Commission Chair Andrew Ferguson pushed back on calls for such an exemption, saying Tuesday he was "highly suspicious" of the request and characterizing it as an attempt to build a moat that would block other competitors from entering the market.

Lehane pointed to the aviation industry as a precedent for safety cooperation among competitors, saying "there are many examples of companies trying to help each other solve safety problems." His argument was that measures aimed at ensuring passenger or technological safety — separate from market competition — fall outside the scope of antitrust law.

The three companies are working through the collaboration to establish an AI standards body, AFP reported Tuesday.

The proposed body is modeled after the Financial Industry Regulatory Authority, which oversees financial firms in the United States. Google Chief Scientist Demis Hassabis had championed the idea.

Lehane plans to meet with members of Congress in Washington to support bipartisan legislation addressing risks posed by AI. "We will support whatever bill can get passed," he said, referring to AI regulatory measures currently advancing in both chambers.

He specifically cited a House bill sponsored by Reps. Jay Obernolte (R-California) and Lori Trahan (D-Massachusetts) on a federal AI governance framework, as well as a Senate bill led by Senate Republican Floor Leader John Thune (R-South Dakota) alongside Sens. Ted Cruz (R-Texas) and Amy Klobuchar (D-Minnesota).

Bloomberg noted that the prospects for passing AI legislation this year appear slim, given that the House is not expected to reconvene until after the midterm elections in early November and that the post-election session is likely to focus on issues such as increasing the defense budget.

President Donald Trump has also been a vocal opponent of AI regulation, dismissing concerns about AI safety as a "scam."

The industry-wide push for safety cooperation gained momentum after Anthropic CEO Dario Amodei called on Saturday for slowing AI development in a personal blog post, with OpenAI CEO Sam Altman and Elon Musk, CEO of SpaceX's AI division, subsequently voicing support for the idea.

Anthropic and Google DeepMind did not respond to requests for comment.


yckim6452@heraldcorp.com