It Turns Out the US Military Really Did Use AI to Blow Up a School Full of Iranian Children

A Pentagon investigation into the February 28 air strikes on the Shajareh Tayyebeh girls' school in Iran concluded the attack, which killed over 150 people including at least 123 children, resulted from outdated intelligence and over-reliance on an AI targeting system, according to new Bloomberg reporting. Anonymous military investigators characterized the double-tap strike as the product of a series of small "errors" compounded by an as-yet unidentified AI targeting system, likely similar to one used to fast-track strikes on over 2,000 Iranian targets within a four-day window. The findings contradict President Donald Trump's July claim that footage of the massacre could have been AI-generated, and follow a UN fact-finding commission report stating there are "reasonable grounds" to conclude the US attacks amount to "war crimes.

Now over half a year into the United States’ war on Iran, it’s becoming abundantly clear that AI has been instrumental to the American campaign. On the home front, AI and its infrastructure is quietly shaping the ideological battlefield: President Donald Trump is spewing out wartime propaganda https://futurism.com/artificial-intelligence/trump-ai-video-fake-attack slopped together by generative AI, as US-owned data centers https://futurism.com/artificial-intelligence/iran-drone-strikes-data-centers-deleting-customer-data and tech company offices https://futurism.com/robots-and-machines/iran-google-microsoft-targets become widely accepted https://www.csis.org/analysis/data-now-front-line-warfare as military targets. AI is also proving instrumental for helping US Central Command choose targets for obliteration via drone strikes or missiles — so much so that an AI system played a pivotal role in perpetuating an American war crime. According to new reporting by Bloomberg https://www.bloomberg.com/graphics/2026-iran-school-attack/ , the Pentagon’s investigation into the February 28th air strikes on the Shajareh Tayyebeh girls’ school — a double-tap strike which killed over 150 people, including at least 123 children — concluded that the attack was the result of outdated intelligence and over-reliance on an AI targeting system. Interviewed by Bloomberg , anonymous military investigators involved in the probe characterized the horrifying attack as the result of a series of small “errors,” rather than one ill-fated mistake. Those errors, stemming from dusty intel, seem to have been compounded by the involvement of an as-yet unidentified AI targeting system https://futurism.com/artificial-intelligence/pentagon-ai-claude-bombing-elementary-school , likely similar to the one used by American military officers to fast-track air strikes https://futurism.com/artificial-intelligence/us-military-elon-musk-grok-bomb-iran on over 2,000 Iranian targets within a four-day window in the early days of the conflict. The latest info directly refutes president Donald Trump’s own narrative around the deadly incident. In July, Trump downplayed the United States’ responsibility for the attack, even suggesting that images and footage of the massacre could have been generated by AI https://futurism.com/artificial-intelligence/trump-images-iran-elementary-school-ai . Yet as a UN fact-finding commission wrote in its report https://news.un.org/en/story/2026/09/1168358 last week, there are “reasonable grounds” to conclude the US attacks amount to “war crimes” — the girls’ school attack being just one prominent example. Whether the full Pentagon investigation will ever be made available to the public is unclear. What is certain is that killer AI is already here — and it doesn’t need to be sentient https://qz.com/sam-altman-ai-safety-pacing-power-concentration-091426 to help humans claim hundreds of innocent lives. More on military AI: These Robot Soldiers Are Getting Downright Terrifying https://futurism.com/robots-and-machines/robot-soldiers-getting-terrifying