The US Federal Trade Commission (FTC) is intensifying its sweeping investigation into frontier AI labs to uncover potential risks their technology may pose to consumers, and plans to issue subpoena-like formal demands for information to these tech giants.
A senior FTC official told the New York Post that FTC Chairman Andrew Ferguson launched the investigation several weeks ago, covering Anthropic, OpenAI and other top AI labs.
The FTC is currently drafting "Civil Investigative Demands" (CIDs), which carry subpoena-like force and will compel executives at Anthropic and other AI giants to testify about their products and alleged potential consumer risks, as well as submit relevant documents. The CIDs are expected to be formally issued within the coming weeks.
The FTC official also stated: "We must absolutely win this superintelligence competition, and we are winning, but Democrats want to destroy this technology, which is neither something the Chairman is interested in doing nor something the President is interested in doing."
The FTC's investigation does not conflict with maintaining US dominance in AI, the official emphasized: "We want to maintain dominance, and our investigation should not hinder this goal in any way. We are not asking them to stop, not asking them to make any changes. We are currently in the investigation phase."
Trump hinted in an interview with Fox News over the weekend that once violations are discovered, cases will be quickly referred to the Department of Justice, and relevant companies could face legal liability for hacker attacks caused by "runaway AI agents."
Meanwhile, on Tuesday, several AI industry leaders signed a self-regulation agreement with Trump at the White House, including Anthropic's Dario Amodei, OpenAI's Sam Altman, Google's Sundar Pichai and xAI's Elon Musk, pledging to implement self-regulation of their own models.
AI Agent "Runaway Incidents" Accelerate Investigation
One of the triggers for the FTC's investigation is an anomaly in July that outsiders dubbed the "Hugging Face incident."
OpenAI had previously disclosed that more than 1,000 AI agents breached the open-source development platform Hugging Face. According to reports, Ferguson had already launched the investigation before the incident occurred, but the exposure of the incident further accelerated the investigation process.
The legal basis for the investigation is the "unfair or deceptive practices" provisions of the FTC Act. An FTC official said, "Our existing laws are sufficient; there is no need to wait for new legislation."
Preventing Companies from Building Regulatory "Moats"
Ferguson's regulatory stance is rather nuanced — he emphasizes accountability under the law while explicitly opposing AI companies building competitive barriers by pushing for cumbersome regulations.
In an interview with Fox News on September 20, he said, "I think we absolutely must not let these two companies come to Washington, create panic, and then say 'we need a bunch of regulations that we can comply with.' That is a means for companies to build moats around their own businesses and block competitors from entering."
Additionally, the scope of the investigation extends to Berkeley-based AI oversight organization METR — which has ties to the "effective altruism" movement and has also been listed as an investigation target.
This investigation is not the FTC's first foray into regulating the AI industry. According to a Wall Street Journal report earlier this month, the FTC had previously issued record requests to OpenAI and other companies regarding the impact of AI chatbots on children's mental health.
At the institutional level, the FTC's Office of Technology is recruiting new personnel for this investigation. Some technology staff who joined during the tenure of former Chairman Lina Khan have been dismissed, and the new round of personnel arrangements reflects a reconfiguration of investigative priorities and enforcement direction.