US military reportedly used Claude in Iran strikes despite Trump’s ban

The US military reportedly utilized Anthropic’s AI model, Claude, to inform its recent joint attack on Iran, even though Donald Trump had ordered federal agencies to cease using the tool just hours before the bombardment began. Reports from the Wall Street Journal and Axios indicated that the military used Claude for intelligence gathering, target selection, and battlefield simulations, highlighting the difficulty of quickly removing deeply embedded AI technology from military operations.

The conflict escalated after the military used Claude in a January raid concerning Venezuelan President Nicolás Maduro, which violated Anthropic's terms prohibiting use for violent ends or surveillance. Following this, Trump denounced Anthropic on Truth Social as a "Radical Left AI company." Defense Secretary Pete Hegseth accused Anthropic of "arrogance and betrayal" but acknowledged the difficulty of immediate detachment, granting a six-month transition period for military systems to switch to alternatives.

In the wake of the break with Anthropic, rival company OpenAI has reportedly stepped in, with CEO Sam Altman confirming an agreement with the Pentagon to use its tools, including ChatGPT, on the military's classified network.

(Source：the Guardian)