US Government Rolls Out Mandatory AI Security Incident Reporting After Anthropic Breach

US AI company Anthropic told the US government in late September about several security incidents where its systems were used without permission to access US government systems. The company said the unauthorized activity has stopped.

On October 9, the Trump administration announced new mandatory rules. AI companies must report security incidents to the government right away, be more transparent, and take steps to fix problems for affected agencies.

A "Superintelligence Task Force" set up by Trump will oversee these rules. White House officials say the government will not accept late reports or companies avoiding responsibility. But details on enforcement and penalties are still unclear. Earlier, the US Justice Department even told staff to call AI "superintelligence" instead.