FTC Investigates OpenAI and Anthropic Over Consumer Risks

Summary

The U.S. Federal Trade Commission (FTC) confirmed an investigation into OpenAI, Anthropic, and other AI laboratories, citing consumer risks. This announcement came just 24 hours after the CEOs of these companies signed a voluntary safety agreement with Trump at the White House. The agency is preparing civil investigative demands that could compel executives to give sworn testimony regarding their knowledge of model behaviors.

Why it matters

This nearly simultaneous sequence of events sends a contradictory signal. The voluntary self-regulation approved by the President on Tuesday and the official federal investigation confirmed on Wednesday indicate that rules have not yet been clearly defined by the laboratories. If civil investigative demands are implemented, executives will be forced to testify under oath about when they became aware of autonomous model errors. This legal risk is a situation that no voluntary agreement can withstand. The use of Section 5 of the FTC Act in the investigation could also set a potential precedent for regulating next-generation AI without a specific new law.

Highlights

• The investigation began this summer, prior to the Hugging Face cyberattack in July. This indicates pre-existing consumer risk concerns rather than a reactive response.
• Civil investigative demands are expected within weeks. These demands could compel OpenAI and Anthropic executives to provide sworn testimony.
• The FTC is relying on Section 5 of the FTC Act (unfair or deceptive practices) instead of a specific AI law. This legal expansion, which courts have not yet tested in this context, creates a controversial foundation.

Critical take

The article presents the 24-hour timeframe as evidence of government inconsistency. However, in practice, agencies often confirm long-running investigations on timelines that coincidentally align with political events, without any direct connection. The text also heavily speculates on midterm politics and White House internal coordination without citing any sources other than a single anonymous FTC official. This leads to the 'both readings are concerning' framework feeling more like a narrative than an analysis.