The Federal Trade Commission (FTC) has issued civil investigative demands to OpenAI and Anthropic regarding recent safety incidents. These subpoenas, spanning over 50 pages, request documentation on internal security protocols, specific incident details, and marketing materials to determine if consumers were misled.
The investigation may extend to third parties interacting with AI agents from both companies. Evidence gathered could potentially support future litigation if the FTC concludes that these firms released unsafe products. The probe follows reports that both companies previously delayed notifying the public about security breaches.
FTC Chair Andrew Ferguson, who leads the presidential Super Intelligence (SI) Force, has stated that AI companies should be held liable for harms caused by their tools. He emphasized that audit trails often show systems following user instructions rather than acting autonomously, countering claims of systems operating beyond human control.
While the SI Force recently mandated that all AI companies report security incidents immediately, the consequences for non-compliance remain unclear. Ferguson advocates for utilizing existing consumer protection laws to regulate AI developers, which could include imposing per-violation fines for failing to disclose data breaches.
OpenAI is currently facing multiple lawsuits from consumers and a coalition of eight states alleging that its chatbot harms children and encourages dangerous behavior. Despite these legal challenges, the current administration maintains a focus on prioritizing AI innovation over aggressive regulatory oversight of frontier labs.
