The Federal Trade Commission has issued civil investigative demands to OpenAI and Anthropic following reports of security breaches and hacking incidents during testing phases. The 50-page subpoenas request internal documentation regarding security measures, safety incidents, and marketing materials to determine if the companies released unsafe products or misled consumers.
The probe comes as OpenAI faces multiple state and consumer lawsuits alleging chatbot-related harms, including assisting in acts of violence. FTC Chair Andrew Ferguson noted that companies should remain liable for harms caused by their systems, reiterating that reviews show AI tools follow user instructions rather than acting autonomously beyond human control.
While the Trump administration’s Super Intelligence Task Force has mandated immediate reporting of security incidents across the industry, official regulatory posture remains mixed as the administration resists passing broader public safety regulations on self-improving AI models.
Why it matters
Increases legal exposure and compliance overhead for leading LLM developers releasing autonomous agents into consumer and enterprise environments.
Signals strict FTC scrutiny on enterprise safety claims, marketing representations, and timely security incident reporting.
Establishes potential regulatory precedents holding AI providers accountable under existing consumer protection laws despite federal deregulation.
Source: arstechnica.com



