Anthropic PBC says it has found evidence that a cell in northern Yemen—an area where Houthi militants operate—used its Claude AI model to support work related to missile and rocket development. The claim centers on the use of Claude within activities described by Anthropic as aiding the development process for weapons systems.
The reporting frames the discovery as part of a broader concern that advanced AI tools can be used by non-state actors for military purposes. Bloomberg focuses on Anthropic’s investigation and its implications for AI misuse risks, while Times of India highlights the allegation more directly as “Houthi” use of AI for missiles based on Anthropic’s findings.
Both outlets point to a growing debate around how generative AI may be exploited in conflict zones, including by groups outside traditional state control. Anthropic’s findings are presented as uncovered by its own analysis of the group’s use of the model, rather than as information attributed to a government investigation or battlefield assessment.
