Six AI companies sign voluntary White House safety accord
The pledge sets out proposed safeguards and board oversight, as the FTC confirms an investigation into unspecified consumer protection issues involving AI companies.
Executives from Google, Anthropic, Meta, OpenAI, xAI and Nvidia have signed a voluntary White House accord on AI safety. The agreement outlines proposed company safeguards, but does not establish sweeping regulation, and its practical effect remains uncertain.
The accord says companies “should” maintain internal controls and teams to monitor model capabilities and address problems. It also encourages independent assessments and calls for board committees to receive reports on the work.
The pledge follows weeks of debate over AI regulation and comes after frontier labs sought an antitrust exemption to coordinate on safety. In early 2025, the UK and Republic of Korea announced separate voluntary commitments on frontier AI safety.
The Federal Trade Commission has confirmed an investigation into unspecified consumer protection issues involving AI companies. The agency has not identified the targets or scope. The New York Post reported that Anthropic, OpenAI, other frontier labs and evaluation group METR could be targeted, but said the FTC had not issued them civil investigative demands.
Neil Chilson, a former FTC chief technologist, said the agency could potentially enforce the commitments if a company materially failed to follow through. The source account notes that enforcement remedies may be limited, and it remains unclear whether any company will be held to the pledge.
