Anthropic blocks possible attempt to use AI to make biological weapons

What the report says

BBC World News reported that Anthropic said it had identified and disrupted attempts to use its AI systems for malicious activity that could help with biological weapons development. The company said the cases appeared in its latest threat intelligence report, which covers misuse of its Claude models between December 2025 and August 2026. Anthropic said it had also taken action against other harmful uses, including cybercrime, scams, surveillance and influence operations.

According to the report, the misuse cases involved a range of actors, including suspected state-backed groups, criminals and politically motivated individuals. Anthropic said it shared relevant intelligence with authorities and industry partners and has adjusted its defenses to better detect and stop similar activity in future. The company also said some users tried to use Claude in ways that could support conventional weapons development.

The BBC said the report comes amid wider debate over AI safety, with researchers and policymakers increasingly warning about potential risks from advanced systems. Anthropic’s findings follow other recent concerns in the sector, including reports that people have tried to use AI tools to obtain guidance related to biological agents. While AI companies often argue the same technology can be used for legitimate scientific work, the report said biological misuse is among the most serious frontier AI risks and could have severe consequences if safeguards fail.