OpenAI Rehearses Post-Catastrophe AI Response

Executives at OpenAI and Anthropic privately rehearse responses to catastrophic AI events, assuming a major incident like a cyberattack on power or banking will occur within six to 12 months. They plan to brief Congress quickly after the Nov. 3 midterms to shape laws, despite fears that an aging Congress might struggle with regulation.