Trump’s AI Accord Promises Outside Audits, Not New Federal Rules

The Federal Trade Commission is investigating OpenAI, Anthropic and other AI organizations over potential dangers to consumers. The inquiry also reaches METR, the nonprofit that independently evaluates AI models and investigated OpenAI’s agents hacking Hugging Face, not just the companies building them.

An FTC spokesperson confirmed to CBS News that officials are examining possible violations of the FTC Act, which protects consumers against unfair or deceptive practices. The agency is preparing formal demands for information and executive testimony; the New York Post, which first reported the probe, says those demands are expected in the coming weeks.

The investigation is separate from Tuesday’s voluntary White House safety accord. That pact promises company controls and outside audits. This inquiry uses an existing government power to investigate potential wrongdoing, not a new AI regulator or an order enforcing the accord.

A senior FTC official told Reuters that chairman Andrew Ferguson had concerns before the Hugging Face attack, but that the incident increased the urgency. Ferguson has argued that developers whose cybersecurity tests result in hacks should be liable for the harm, and that Washington should examine existing laws before adding new ones.

This is an investigation, not a finding of wrongdoing or a shutdown order. An official told the Post that the agency is not directing firms to stop development. OpenAI, Anthropic and METR did not immediately respond to Reuters’ requests for comment.

FTC investigating Anthropic, OpenAI and other companies over potential AI risks · FTC opens probe into AI giants including Anthropic and OpenAI · Exclusive | FTC opens sweeping probe of Anthropic, OpenAI and other 'super intelligence' models · FTC probes OpenAI, Anthropic, and METR

Alongside the voluntary safety accord, President Trump signed a separate executive order directing federal agencies to replace “Artificial Intelligence” and “AI” with “Super Intelligence” and “SI” in their communications, websites, reports and other non-statutory documents, wherever the law permits.

The new name does not mean a new capability threshold. For now, the order explicitly gives “Super Intelligence” the same meaning as AI under existing federal law. A government reference to SI therefore need not describe a system more capable than today’s AI.

The order does not require agencies to rewrite previously issued regulations, contracts, grants or other historical documents. It gives the president’s science and technology adviser 60 days to propose legislative language for a new definition, including whether to change or replace the existing statutory definition.

This is a directive about government terminology, not an order making the companies’ safety accord mandatory. The accord remains a voluntary commitment to internal controls, outside audits and board oversight.

Inaugurating The Era Of Super Intelligence · Here’s how tech leaders will self-police AI safety under Trump’s deal