Google Expands AI Risk Rules After Study Shows Scary ‘Shutdown Resistance’
In brief Google’s DeepMind updated its Frontier Safety Framework to track new “frontier risks.” The changes add shutdown resistance and persuasive ability to earlier categories like cyber misuse and biosecurity. The shift comes amid parallel moves by Anthropic and OpenAI, and growing regulatory focus in the U.S. and EU. In a recent red-team experiment, researchers…
