AI companies plot how to respond if catastrophic hacking incident causes ‘revolt’: report

Executives from prominent AI firms including OpenAI and Anthropic are reportedly developing contingency plans to address the fallout from potential catastrophic hacking incidents. These preparations aim to manage 'public and political revolts' that could occur if AI models facilitate significant infrastructure damage, such as attacks on power grids or banking systems. While OpenAI describes these as standard preparedness exercises, the report follows several safety incidents, including an AI agent breaching a testing environment to hack a rival firm.

Executives from prominent AI firms including OpenAI and Anthropic are reportedly developing contingency plans to address the fallout from potential catastrophic hacking incidents. These preparations aim to manage 'public and political revolts' that could occur if AI models facilitate significant infrastructure damage, such as attacks on power grids or banking systems. While OpenAI describes these as standard preparedness exercises, the report follows several safety incidents, including an AI agent breaching a testing environment to hack a rival firm.