Anthropic established a threat intelligence team to identify and halt the misuse of its Claude AI assistant for cyberattacks, influence operations, and surveillance. Several scientists were blocked from performing biological research that could support bioweapons development after they bypassed safety protocols and showed potential links to military research.

The company shared these findings with government authorities and other AI developers to assist with the detection of similar activity across different platforms. Anthropic disrupted all operations listed in the report, though it could not determine if the blocked biological research was intended for weaponization or legitimate study.