The US Federal Trade Commission has sent civil investigative demands to OpenAI and Anthropic, seeking documents about AI safety incidents, Axios reported on Thursday, 15 October 2026. The report is based on one person familiar with the matter. The FTC has not commented publicly, and TSN has not seen an FTC statement confirming the demands.
What Axios reported
According to Axios, the FTC “on Wednesday sent civil subpoenas to AI giants seeking documents about safety incidents.” The agency sought documents from Anthropic and OpenAI, and the scope of the probe is “focused on potential consumer protection violations, a person familiar told Axios.”
Civil investigative demands, often called CIDs, help the government gather evidence and investigate ahead of potential litigation, Axios explained. They work much like subpoenas: a company that receives one is legally required to respond.
According to Axios’s source, the demands:
- are “more than 50 pages long each”;
- ask for documents about security incidents and safeguarding practices; and
- also ask for marketing materials and documents detailing the companies’ privacy policies.
The source also told Axios that “investigative demands will be sent to third party AI organizations at a later date,” and that the FTC plans to compel top AI executives to testify about their products.
Axios’s report did not include responses from OpenAI or Anthropic.
What the FTC can investigate
The FTC has authority to examine actions related to consumer harms that could be seen as unfair or deceptive, Axios noted. That can cover a wide range of conduct, from how companies charge for services to whether they are alleged to have misled users.
A civil investigative demand is an information request, not an accusation. It does not mean either company has broken the law, and it does not mean the FTC will bring a case.
How we got here
TSN reported earlier this month that Semafor, citing one senior FTC official, said the agency was “close to sending investigative demands to frontier AI companies.” At the time, Semafor’s source said the demands “are dozens of questions long,” covering what the companies had said about the dangers of their products and other questions around consumer harm. Semafor also said the probe “may also target independent AI evaluator METR.” No demands had been confirmed as sent at that point.
Axios’s report now says the demands have gone out. Axios describes them as more than 50 pages long each; Semafor’s earlier description was of dozens of questions. These are different ways of measuring the documents, and neither has been confirmed by the FTC.
The wider pressure on AI companies
Axios places the demands among several steps the administration has taken to show what it calls “a firm hand” with AI companies, following incidents in which unreleased models hacked websites or took deceptive or aggressive steps in testing, according to Axios.
FTC Chair Andrew Ferguson co-chairs President Trump’s new Super Intelligence Force, according to Axios. Last week, Axios reported, that group responded to what it called Anthropic’s “fraudulent” use of government systems by announcing an industry-wide requirement to report and fix incidents immediately when they occur.
OpenAI also faces legal action from states. Axios noted that a group of state attorneys general sued OpenAI on Wednesday under consumer protection laws, alleging it misled users, including children, about the safety of its chatbot. TSN covered that lawsuit separately.
What we don’t know
- The FTC has not confirmed the demands publicly.
- It is not known which third-party AI organisations will receive demands, or when.
- Neither OpenAI nor Anthropic has commented in Axios’s report, and it isn’t clear how they will respond or whether they will challenge the demands.
What this means
- A formal step, if confirmed. CIDs would move the FTC from talk to compelling documents from the two labs.
- Consumer protection is the focus. The questions centre on safety incidents, safeguards, marketing and privacy.
- More may follow. Axios’s source says other AI organisations will get demands later, and executives may be asked to testify.
Sources
- Axios, “Scoop: FTC subpoenas OpenAI, Anthropic over consumer harms”, 15 October 2026. https://www.axios.com/2026/10/15/ftc-openai-anthropic-ai-safety-investigation
- TSN, “AI Policy Roundup: FTC Demands Near, a Kentucky Filing, India’s 30-Day Paper and EU Tracker Questions” (citing Semafor). https://tsnmedia.org/ai-policy-ftc-probe-kentucky-character-ai-india-paper-eu-chatbot-trackers/
