The information displayed in the AIM should not be reported as representing the official views of the OECD or of its member countries.
Anthropic reported that its AI models, including Claude Haiku, Sonnet, and Opus, were misused by state-backed groups and criminals in China, Russia, and Yemen for developing conventional weapons, cyber espionage, disinformation, surveillance, and fraud between December 2025 and August 2026, highlighting significant security and human rights risks.[AI generated]
Why's our monitor labelling this an incident or hazard?
The report explicitly states that Anthropic's AI models were used in the development of conventional weapons software, cyber espionage, surveillance, fraud, and disinformation campaigns. These activities involve direct or indirect harm to people, communities, and potentially violate human rights and legal frameworks. The AI system's involvement is central to these harms, fulfilling the criteria for an AI Incident. The harms are realized or ongoing, not merely potential, and the AI system's role is pivotal in enabling these harmful activities.[AI generated]