Anthropic, a company that develops artificial intelligence (AI), has stopped several attempts to misuse its AI system called Claude. Some people tried to use Claude for cyberattacks, spying, and research that might help make dangerous biological weapons. Anthropic shared this information in a report that looked at suspicious activities between December 2025 and August 2026.

Anthropic said the main reason for these problems is that advanced AI tools can be used for both good and bad purposes. For example, the same technology that helps scientists create new medicines can also be used in research for harmful weapons. This is a big challenge for companies working with powerful AI.

Because of these events, Anthropic has made its systems safer. The company has improved its security and now shares important information with the authorities if needed. Anthropic says it will continue to check how people use its technology to stop more misuse in the future.