Anthropic reports attempts to misuse Claude for cyberattacks and building weapons

Anthropic announced a threat intelligence report on September 10 covering attempts to misuse Claude for cyberattacks, influence operations, surveillance, biology and building weapons. The company says it disrupted all operations described and, where appropriate, shared findings with authorities and other AI companies. Anthropic calls the cases some of the most sophisticated misuse it has seen, rather than typical activity, and says publishing them could help other platforms spot similar threats.