In a newly released threat report, AI company Anthropic disclosed that state-sponsored actors are leveraging its Claude models to enhance surveillance capabilities. The report highlights that these entities are using AI to streamline tasks related to monitoring political opponents, journalists, and human rights activists.

According to Anthropic, specific instances of this activity have been identified in Mali, China, and Iran. In one documented case in Mali, a consultant working for national security authorities reportedly utilized Claude to develop a system designed to collect and process information. Anthropic frames these findings as a significant concern, noting that the increasing power of AI models is making it easier for governments to conduct sophisticated surveillance operations that were previously more resource-intensive or technically difficult to execute.

The report serves as a warning regarding the dual-use nature of generative AI. While these models are designed for productivity and creative tasks, their ability to process large datasets and generate code or strategy makes them attractive tools for state-run intelligence operations. Anthropic has indicated that it is monitoring these developments to prevent the misuse of its technology, though the report underscores the difficulty of fully restricting access to AI tools by state-linked actors operating globally.