Anthropic Claude misuse cases included working scientists whose activity could support biological weapons development, prompting the AI company to ban their accounts, according to a threat report released Thursday.
Anthropic said it identified five biological-risk cases but could not establish that the scientists intended harm. It also said the incidents did not prove that Claude could be used to create biological weapons in the real world.
The company said it incorporated findings from the investigations into its safeguards, enforcement systems and threat-intelligence processes.
Anthropic described biological misuse as one of the most serious risks associated with frontier AI models. It said newer systems, including Claude Fable 5, impose stronger restrictions on a broad range of dual-use biological research queries.
Read: Jacob Coxon Resigns, Says AI 10% Chance of ‘Killing All Humans’
The other alleged misuse involves propaganda, surveillance and military targeting. Anthropic said it removed three Iranian state-aligned accounts that used Claude for influence operations.
The company said one Iran-linked actor also used the model to develop targeting recommendations involving U.S. naval forces and software for a domestic surveillance platform.
Read: Anthropic Launches Project Glasswing to Fight AI Cyber Threats
Separately, Anthropic said three accounts aligned with a Chinese municipal security service used Claude for surveillance and transnational repression, including profiling activists and organisations abroad.
