US artificial intelligence company Anthropic has identified and disrupted multiple attempts to misuse its Claude AI models, including five cases involving potential biological weapons development. The company's first threat intelligence report of the year, covering activity between December 2025 and August 2026, also documented misuse by actors linked to Russian cyber espionage, an Iranian propaganda institution, and China-based hacking groups, as well as cases involving conventional weapons software, surveillance tools targeting dissidents, and online scams. The findings add to mounting concern within the AI industry about frontier model safety, with Anthropic warning that biological misuse in particular "could have catastrophic consequences" and saying it has shared relevant intelligence with authorities and industry partners.