AI Is Arming America’s Enemies
Anthropic’s own threat report shows hostile governments and militias turning American AI into a force multiplier. A Houthi-linked engineering cell used Claude to help develop guidance systems for rockets and a missile with a range exceeding 2,000 kilometers, run simulations and diagnose a failed test launch. The model blocked some requests, yet engineers evaded other safeguards by dividing their work across multiple sessions. Iranian actors used the technology to build propaganda networks, profile targets and track U.S. naval forces, while Chinese and Russian operations applied it to cyberespionage. AI is lowering the cost of capabilities that once required teams of engineers, analysts and linguists—giving weaker states and terrorist groups a faster path to weapons, surveillance and disruption.
