US strike on Minab School resulted from incorrect AI-Generated Intelligence, report

Tehran: A reported Pentagon investigation has shed new light on the US strike that hit a girls’ school in Iran’s southern city of Minab, pointing to a chain of failures involving outdated intelligence, an accelerated targeting process and heavy reliance on artificial intelligence.

According to a Bloomberg investigation, officials familiar with the Pentagon probe said the Shajarah Tayyebeh school was incorrectly classified in US military intelligence databases as an Islamic Revolutionary Guard Corps (IRGC) facility. The full Pentagon investigation has not been publicly released, so some details remain based on accounts from officials involved in the inquiry.

The school was located on land that had previously formed part of a military compound. However, satellite imagery showed that the site had undergone significant changes, including separate walls and entrances, a playground and other visible signs of civilian use.

According to the report, some US intelligence analysts had identified changes at the location as early as 2019. However, the information was reportedly recorded in a system that was not connected to the main intelligence database used for targeting. As a result, the older military classification remained in the targeting system.

How AI entered the targeting process

The report says the AI-enabled Maven Smart System, developed by Palantir Technologies, was used to process and coordinate large amounts of military intelligence.

Officials familiar with the investigation said some personnel relied heavily on the system and expected it to identify inconsistencies or outdated information. However, the available reporting does not indicate that AI independently decided to attack the school. Rather, the technology was part of a broader human-led targeting process operating with flawed underlying information.

The targeting process was also reportedly compressed because US forces were preparing a large-scale opening assault against Iranian targets. More than 1,000 targets were reportedly struck during the first 24 hours, increasing pressure on military planners.

The investigation also examined reduced staffing for civilian-harm mitigation teams. According to officials cited by Bloomberg, the US Central Command team had been reduced from 10 personnel to one, and no member reviewed the Minab site before the strike.

A devastating outcome

Two US Tomahawk missiles struck the school on February 28, 2026, killing more than 150 people, including at least 123 children, according to the Bloomberg investigation.

The incident has since become a major case in discussions about the risks of using AI-assisted systems in military targeting.

Palantir has disputed suggestions that its software was responsible for the intelligence failure, saying the company does not control the underlying data used by the system.

The reported findings have nevertheless raised broader questions about whether increasingly automated military systems can unintentionally accelerate flawed intelligence instead of detecting it.

As the Pentagon investigation remains unpublished, the precise distribution of responsibility among intelligence databases, human reviewers, AI-assisted technology and military command decisions has yet to be fully established.