FTC Investigates OpenAI, Anthropic and Other AI Companies Over Product Risks

Illustration for: FTC Investigates OpenAI, Anthropic and Other AI Companies Over Product Risks
AI-generated illustration. Visual interpretation does not represent real individuals or scenes.

THE BARE STORY

The Federal Trade Commission has opened an investigation into OpenAI, Anthropic and other artificial intelligence companies over potential risks their products may pose to consumers, the agency confirmed Wednesday.

The FTC said it began the inquiry this summer. An agency spokesperson did not identify the other companies included. The regulator is examining whether company practices comply with the Federal Trade Commission Act, according to the FTC.

The agency plans to seek detailed information from the companies and from the nonprofit research group METR, according to the FTC. Regulators are also preparing civil investigative demands for AI executives to provide testimony about their products.

The investigation follows a White House meeting Tuesday involving executives from companies including Anthropic, OpenAI, Meta, Nvidia and SpaceX. Participants signed a voluntary safety accord, while President Donald Trump said his administration would not slow AI development.

Same Facts. Different Perspectives.

Three AI models. Three viewpoints. One factual foundation.

Start with what this actually is: an FTC inquiry under existing law, not a new AI statute, and not an enforcement action yet. That distinction matters. Congress has not passed AI-specific consumer protection legislation, so the agency is doing the pragmatic thing — using the tools it already has, the FTC Act's unfair-and-deceptive-practices authority, to ask whether AI companies' claims and safeguards hold up. That's institutional competence working as intended: adapt existing authority rather than wait years for legislation that may never arrive.

The decision to pull in METR is the most reassuring detail here. Consumer protection cases involving opaque, fast-moving technology routinely fail when regulators lack the technical grounding to know what they're actually looking at. Bringing in an independent AI-evaluation research group suggests the FTC is trying to build an evidentiary record rather than run a political exercise. That's the right instinct, though it's still just an instinct at this stage — a civil investigative demand is not a finding, and 'examining compliance' is not the same as 'violations found.'

What's genuinely striking is the timing. One day, AI executives sign a voluntary safety accord at the White House while the President insists nothing will slow AI development down. The next day, the FTC quietly confirms it's been building a formal investigation since summer. That sequence tells you something real: voluntary industry pledges are not a substitute for enforcement, and the federal government itself doesn't appear to be relying on them as one. Whether that's healthy separation between an independent agency and White House messaging, or simply incoherent signaling to an industry trying to figure out what Washington actually wants, is a fair question — probably some of both. Industries don't plan well when the regulatory posture and the political posture point in different directions.

The open question is scope. 'Product risks to consumers' could mean deceptive capability claims, inadequate safety guardrails around vulnerable users, data practices, or something narrower. The FTC Act is a consumer-protection statute, not an AI-safety charter — it's built for fraud and deception, not necessarily for the harder questions about model behavior at scale. If regulators try to stretch decades-old unfair-practices doctrine to cover novel technical risk, they may find the legal fit awkward. That's not a reason to avoid the inquiry; it's a reason to be skeptical of dramatic near-term outcomes.

How it may affect me

Nothing changes for users of ChatGPT, Claude, or other AI products today — this is a fact-finding investigation, not a rule change or penalty. Realistically, any consequence is measured in years, not months: civil investigative demands and executive testimony take time to produce, and any enforcement action would come well after that.

Over a longer horizon, if the FTC finds evidence of deceptive marketing or inadequate safety disclosures, expect clearer labeling of AI capabilities and limitations, possible changes to how these products are marketed to consumers, and potentially fines or consent decrees — the FTC's usual toolkit. For companies, the immediate cost is compliance: legal resources, documentation, and executive time responding to demands, which can slow product decisions even before any finding of wrongdoing.

For ordinary people, the more meaningful effect may be less direct — a signal that oversight of AI safety claims exists somewhere in government, even as the administration simultaneously promises not to slow the industry down. Whether that oversight produces real consumer protections or gets outpaced by how fast these products are shipped and used is genuinely uncertain, and will depend heavily on how seriously the investigation is pursued once it moves beyond information requests.

Read the story at

Note: All TheBareNews content is AI-generated. For additional context, reporting, and updates, you are invited to explore the news outlets linked above.

Stories You May Have Missed