New Delhi: The US military reportedly relied on Anthropic’s Claude AI model during its major strikes on Iran, even after President Donald Trump ordered federal agencies to stop using the company’s technology just hours earlier. Military commands are said to have used the AI for intelligence processing, identifying targets and running battlefield scenarios as part of the joint US-Israel offensive.
Trump’s public move to sever ties with Anthropic came amid a bitter dispute over military use of AI. The president criticised the company on social media and directed an immediate halt to its AI tools in government work, a ban that has sparked fierce debate about ethics and national security.
AI on the battlefield despite ban
According to the Wall Street Journal, US Central Command units deployed Claude’s capabilities during the operation on Iran, underscoring how deeply integrated the technology has become in defence systems. The model supported tasks including data analysis and combat planning that were central to the strike’s execution.
The use of Claude came hours after Trump’s directive, revealing the difficulty of rapidly removing an AI tool already embedded across classified military networks. Officials familiar with the situation told The Wall Street Journal that its continued use reflects both reliance on the technology and the administrative challenge of an immediate cutoff.
Trump-Anthropic fallout
Anthropic has been severely criticised by Trump. He described the company as a radical left AI firm and the management of the company as having no understanding of the real-world needs, and it was putting national security at risk. The prohibition happened after a high-profile dispute in which the US military used Claude to raid Venezuelan President Nicolas Maduro, going against the terms of use in the company.
The defence secretary, Pete Hegseth, also condemned the company, claiming that its attitude was one of arrogance and betrayal and making a demand that the company make its AI models available to everyone in ways that complied with the law. Hegseth has also made it possible to have a maximum of six months to transition to replace Claude as new systems are implemented.
Following the split with Anthropic, competitor machine intelligence company OpenAI has entered into a deal with the Pentagon to supply its own technology, such as ChatGPT, to operate on classified military networks. The deal was confirmed by the CEO of OpenAI, Sam Altman, and its goal is to become the next backbone of defence AI.