The US-Israel strikes on Iran showed that advanced commercial AI tools are no longer fringe experiments; they’re woven into real battlefield planning in ways that are hard to disentangle.
The operation also highlighted a broader trend in modern warfare: once AI systems are integrated into military operational networks, they can’t simply be switched off overnight.
According to reporting by the Wall Street Journal, US military commands continued using Anthropic’s Claude AI model in planning and intelligence support for the strikes, even as the Trump administration had directed federal agencies to begin phasing out its use.
The Pentagon’s arrangements allow up to six months’ transition because the technology is deeply embedded in classified systems.
On Truth Social, Trump announced, “I am directing EVERY Federal Agency in the United States Government to IMMEDIATELY CEASE all use of Anthropic’s technology. We don’t need it, we don’t want it, and will not do business with them again!”
The following day, the US and Israel launched Operation Epic Fury, a coordinated campaign targeting Iranian leadership and military infrastructure after diplomatic efforts to limit Tehran’s nuclear program collapsed.
During the strike, Claude was not used to autonomously select targets or control weapons systems. Instead, it supported intelligence analysis, operational planning, and command-level assessment.
It wasn’t the first operational use of Claude, which had previously supported classified missions, including the capture of Venezuelan leader Nicolás Maduro.
As additional systems such as xAI’s Grok and OpenAI models enter classified networks, AI developers are increasingly functioning as defense infrastructure providers rather than external technology vendors.
