A deadly Tomahawk strike on an Iranian civilian school and growing reliance on AI targeting tools have forced the Pentagon to confront the dangerous intersection of automation bias and military decision-making.
The US military’s experiment with AI-powered intelligence is producing results that look less like a technological breakthrough and more like a cautionary tale. A Tomahawk missile strike on February 28, 2026, killed an estimated 150 to 175 civilians at a school in Minab, Iran, after outdated intelligence was fed through AI-assisted targeting systems without adequate human vetting.
What went wrong in Minab #
The strike relied on stale data that had not been properly validated before the launch order was given. Subsequent investigations found that the errors were primarily attributable to human decision-making, not AI hallucination in the technical sense. Operators chose to disregard database warnings and proceeded with information that no longer reflected conditions on the ground.
US Central Command has been at the center of the military’s AI adoption push. CENTCOM Commander Admiral Brad Cooper acknowledged in mid-2026 that AI tools are being used for rapid data processing across operations, while maintaining that human decisions remain the ultimate authority.
The scale of AI targeting #
Palantir’s Maven Smart System, one of the primary platforms integrated into CENTCOM operations, reportedly generated around 1,000 targeting recommendations within the first 24 hours of a single campaign phase.
CENTCOM has been working with Anthropic’s Claude model alongside Palantir’s data infrastructure to handle the firehose of intelligence data that modern operations generate. Defence analysts have cautioned that AI systems operating on synthetic or corrupted data can produce outputs that look authoritative but are fundamentally disconnected from reality.
AI, tech, and the markets they move—in one daily briefing.
Daily. Free. Join 34,000+ readers across crypto, finance, and policy.
The adversary is watching #
Anthropic published reports on September 10 and 11, 2026, revealing that Iranian actors had been using similar AI systems to develop targeting strategies against US Navy assets. Iranian operators reportedly created AI-assisted targeting handbooks specifically designed to identify and strike American naval platforms.
Where this leaves the Pentagon #
The Pentagon now faces intensifying scrutiny over its AI integration framework. The existing policy mandates human oversight at every stage of the targeting process, but the Minab incident demonstrates that having a human in the loop is meaningless if that human defaults to the machine’s judgment.
Some defense researchers have proposed mandatory “red team” verification layers, where separate AI systems or analyst teams are tasked with challenging targeting recommendations before they reach decision-makers. Others argue that the real fix is cultural: training operators to treat AI outputs as hypotheses rather than conclusions.
The Minab strike didn’t fail because the AI was broken. It failed because the data was stale and the humans trusted the system more than the warnings it was already generating.
Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our