Anthropic Blocks Attempts to Use Claude AI for Bioweapons, Cyberattacks and Propaganda

Anthropic blocked attempts to use its AI models for bioweapons research, cyberattacks, and propaganda between December 2025 and August 2026. One case involved drafting a grant application for gain-of-function research on chikungunya virus. State-sponsored groups, spyware vendors, and criminal hackers were among the misusers.