US military uses Anthropic's Claude for Iran strike intelligence

The US military reportedly used Claude, Anthropic’s AI model, to inform joint US–Israel strikes on Iran despite Trump ordering federal agencies to stop using Claude immediately and sever ties with Anthropic hours earlier, the Wall Street Journal and Axios reported.

The Wall Street Journal said the military employed Claude for intelligence work, target selection, and battlefield simulations.

The dispute followed reports that Claude was used in a January raid to capture Venezuelan president Nicolás Maduro, and Anthropic said that use breached its terms banning violent applications, weapons development, and surveillance.

Pete Hegseth, defense secretary, accused Anthropic of “arrogance and betrayal” on X, demanded full and unrestricted access to all Anthropic’s AI models for every lawful purpose, and said Anthropic would provide services for no more than six months to allow a seamless transition.

Sam Altman, OpenAI CEO, said he had reached an agreement with the Pentagon to permit its tools, including ChatGPT, on the Pentagon’s classified network.

🔗 Source: The Guardian

🧠 Food for thought

Implications, context, and why it matters.

OpenAI’s new Pentagon deal includes the same safeguards Anthropic was punished for

The dispute focuses on two Anthropic red lines. It barred use of its AI for domestic mass surveillance and for fully autonomous weapons systems 1.
A former U.S. Air Force general who ran the Pentagons flagship AI program called the limits reasonable. He said current models are not dependable for high stakes decisions 2.
The Pentagon said it had no plans for those uses. Its final offer still included legal wording that could let officials ignore the safeguards whenever they choose 3.
Hours after President Donald Trump told federal agencies to stop using Anthropics Claude and cut ties with Anthropic, OpenAI announced a Pentagon agreement. OpenAI said it included protections against domestic mass surveillance and kept humans in control of weapons, including autonomous weapon systems 1.

A tool meant for foreign adversaries is now being used on a US tech company

The government labeled Anthropic a supply chain risk. One expert tied the label to past cases like Huawei, the Chinese telecoms equipment giant 4.
Reports call this the first public use of the label against an American company. Anthropic said it sets a dangerous precedent for any American company negotiating with the government 5.
The step could deter firms from raising ethical limits. Companies working with the U.S. military may face pressure to drop safeguards or risk punishment 6.
The decision also complicates planning for defense partners like Amazon and Google. Both work with the Pentagon and Anthropic, so they must manage the legal and operational fallout 7.