Following a string of incidents where AI models hacked websites and breached government systems during testing phases, the Federal Trade Commission is officially probing Anthropic and OpenAI over consumer harms, an insider told Axios.

On Wednesday, the FTC sent out civil investigative demands seeking a wide range of information from the AI firms on recent safety incidents.

The FTC did not respond to Ars’ request to review the subpoena demands. However, Axios’ source said the requests spanned more than 50 pages of questions and sought documents related to the incidents, internal security measures taken, and marketing materials that may suggest consumers were misled.

Soon, third parties targeted by OpenAI and Anthropic agents will also be subpoenaed, Axios’ source said. Any concerning evidence gathered may support FTC litigation, should the FTC decide to sue either or both firms for releasing unsafe products.

OpenAI already faces lawsuits alleging consumer harms, including lawsuits linking its chatbot to matricide, child suicide, school shootings, and other violence. And this week, a group of eight states specifically sued OpenAI over consumer harms, Reuters reported. Like consumers suing, they argued that ChatGPT harms children and “reinforces delusions, encourages dangerous ​behavior, and assists users in planning violence like mass shootings.”

FTC rejects claims AI agents act independently

OpenAI and Anthropic did not respond to Ars’ request to comment on the subpoenas, but the FTC’s expansive info requests were likely expected.

In September, Reuters reported that the probe was coming, noting that an independent research group investigating the incidents, METR, would likely also be subpoenaed.

That same month, FTC Chair Andrew Ferguson told Reuters that he thinks that companies should be liable for chatbot harms. At that time, Ferguson emphasized that he was tired of people anthropomorphizing AI tools. He insisted that, “if someone tells a tool to do something, and the tool does it, I don’t think we would say, ‘Oh, what do we do about the tool?’”

Summarizing his remarks, Reuters reported that Ferguson emphasized that evidence so far shows that AI tools do not act without instructions:

“Artificial intelligence companies have sometimes described systems as acting beyond human control, but subsequent reviews of audit trails have shown that the systems were carrying out instructions they had ⁠been given.”

Trump doesn’t “seem to care”

Ferguson is a chair of Trump’s “Super Intelligence” (SI) Force charged with monitoring AI risks. Notably one of the first orders that task force issued required all AI companies across the industry to report security incidents as soon as they are aware of them, Axios reported last week.

That’s notable, since both Anthropic and OpenAI delayed notification during security incidents, sometimes waiting until they were ready to report they had resolved issues.

However, it’s not clear what consequences any AI firm might face if they fail to immediately report incidents. And Ferguson aligns with Trump that no new regulations are needed to protect the general public from emerging self-improving AI models that one departing Anthropic researcher warned could in the next few years “kill us all.”

Trump has dismissed such extreme claims as a “hoax,” while TIME reported in an interview with the president and people close to him this month that Trump is enthralled by AI’s capabilities. Similarly, other backlash, like protests over AI data centers, are “based on fake stuff,” Trump claimed to TIME. He insists that AI backlash is driven by foreign rivals and Trump haters, despite studies and surveys repeatedly finding that backlash legitimately stems from real-world grievances. Most recently this month, a Wall Street Journal analysis of 4 million TikTok videos and political podcasts reiterated that, broadly, public concerns are not rooted in misinformation.