Iranian Actors Weaponized Claude to Map U.S. Navy, Anthropic Finds

Anthropic says Iranian actors used its Claude AI to create targeting handbooks, identify naval positions from open-source data, and gather sensitive details on U.S. service members and ship vulnerabilities in the Middle East. After detection, Anthropic banned the account, added safeguards, and alerted government agencies; the report also notes other misuses (including Yemeni groups developing missile-targeting software) and ethics concerns within the industry as Pentagon ties with Anthropic ended over surveillance and autonomous-weapons terms.