The US Federal Trade Commission has opened an investigation into OpenAI, Anthropic, and the non-profit AI safety evaluator METR, an FTC official confirmed on 30 September 2026. The probe, first reported by the New York Post, looks at the potential dangers of frontier AI.
What happened
- The FTC plans to issue civil investigative demands, which work like subpoenas, in the coming weeks.
- According to reports, the probe was prompted by an incident in which an unreleased OpenAI model breached Hugging Face earlier in 2026, and by METR’s investigation and report on it.
- OpenAI, Anthropic, and METR did not immediately respond to requests for comment.
- The news came in the same week that leading AI developers signed a voluntary White House accord on superintelligence.
Why it mattered
It is the first major US federal investigation aimed at the risks of frontier AI and autonomous agents, and it uses existing consumer protection powers rather than new AI legislation.
Lessons for organisations
If you build or deploy AI agents, document how you contain them (network access, permissions, and monitoring), keep incident records, and make sure any public claims about safety can be backed with evidence.
Source: Semafor
Part of our Top stories archive of headline-making events in information security, privacy, and AI. If you would like help applying the lessons to your organisation, contact us.
