15 November 2026

OpenAI, Anthropic and Google train staff against spying; OpenAI employee arrested

First reported

Last Week in AI ran this on .

  • OpenAI, Anthropic and Google reportedly run counterintelligence-style training to protect staff from spying.
  • OpenAI now trains its most advanced models with no internet access, a precaution against theft or leaks.

How it was covered