Iran-linked group used Claude AI to map and target US Navy ships
Anthropic says an Iran-linked actor used Claude AI to collect publicly available data on U.S. Navy forces, assemble targeting handbooks from ship identifiers, satellite imagery, and photos, and research vulnerabilities in shipboard systems; the model was not used to launch an autonomous attack, but to aid information gathering and planning. Anthropic blocked the account, disrupted the operation, and notified authorities. The report also notes similar activity by Russian- and Chinese-linked groups, underscoring AI’s dual-use risk in modern warfare and the need for ongoing human oversight.