The Shajareh Tayyebeh primary school in Minab, a city in southern Iran’s Hormozgan province, was struck three times on the morning of 28 February. It was a Saturday, a regular school day in Iran. Around 170 girls aged between seven and twelve were at their desks when the first missile hit.
The school principal moved a group of survivors to a prayer hall and called parents to come and collect their children. A second strike hit that area before most parents could arrive. Between 165 and 180 people died.
At least 95 more were injured. 69 schoolgirls have still not been identified and their remains are undergoing DNA testing. It is the deadliest civilian strike of the war so far.
What Investigators Found
A preliminary US assessment suggests the United States was likely responsible for the strike and may have hit the school in error, possibly due to the use of dated intelligence which wrongly identified the area as still part of an Iranian military installation.
The school had a complicated history. The building was initially walled within the IRGC compound in 2013 but was walled off by September 2016, becoming a clearly defined civilian institution for more than ten years. The Guardian found no indication it served any military purpose at the time of the strike, and determined that the compound buildings immediately adjacent to the school were a medical clinic and a pharmacy.
Analysis of strike locations revealed an unusual pattern: missiles hit the military base and the school, but bypassed the clinic located between the two without touching it, strongly indicating that the executing party was operating with coordinates and maps that distinguished between the different facilities.
This Week in Worcester, citing a Department of Justice appointee speaking anonymously, reported that the working theory inside the investigation is that an AI targeting programme included the school’s coordinates based on archived intelligence that pre-dated the 2016 separation.
A Department of Defense logistics programmer told the same outlet that the Pentagon had rapidly scaled up its use of a Claude-based system over the past year, integrating it into many core operational decisions.
Two sources familiar with the matter confirmed to CBS News that the US military used Claude in its strikes on Iran, though it is not clear exactly how the model was deployed.
The Anthropic Dimension
The AI system at the centre of the targeting question is Claude, made by Anthropic. Anthropic signed a $200 million contract with the Department of Defense and was the first AI lab to deploy its technology across the agency’s classified networks. The relationship has since collapsed in a dispute that now sits in federal court. The Pentagon wanted Anthropic to grant it unfettered access to Claude across all lawful purposes.
Anthropic refused, insisting on two red lines: that Claude would not be used for mass surveillance of US citizens, and that it would not be used for fully autonomous weapons. The Trump administration cancelled the company’s deals and designated it a supply-chain risk, a label historically reserved for companies tied to foreign adversaries, with a six-month phase-out period. OpenAI announced its own Pentagon deal within hours.
Anthropic sued the Pentagon on Monday, asking the court to vacate the supply chain risk designation. The company said the actions were unprecedented and unlawful and that they could reduce its 2026 revenue by multiple billions of dollars. Dozens of researchers at OpenAI and Google DeepMind filed an amicus brief in their personal capacities supporting Anthropic, arguing the designation could harm US competitiveness and hamper public debate about AI risks.
The day after the Pentagon’s blacklisting, Claude surpassed ChatGPT in the iPhone App Store for the first time. More than a million people are signing up for Claude every day, according to Anthropic.
The Accountability Gap
The Minab strike forces a question that AI ethics discussions have circled for years without resolving: when an AI system contributes to a decision that kills civilians, who is responsible. Human Rights Watch has called for a thorough independent investigation into the strike on Shajareh Tayyebeh Primary School to determine whether war crimes were committed.
Defense Secretary Pete Hegseth said the military is investigating and does not target civilians. The chairman of the Joint Chiefs of Staff, General Dan Caine, said at a Pentagon briefing that US forces were operating along the southeastern coast in an area that included Minab on the day of the strikes.
The specific chain of authorisation matters enormously and remains unclear. The This Week in Worcester source described uncertainty over “the logic behind the launch and the mechanics of who authorised it.”
If an AI system flagged the school’s coordinates as a valid military target based on outdated data, and that flag was passed to a human operator who approved the strike without independent verification, the question of liability sits precisely at the seam between human judgment and algorithmic recommendation.
This is exactly the scenario Anthropic’s red lines were designed to prevent, and exactly the scenario that unfolded while those red lines were still nominally in force.
What the Market Now Understands
The AI sector absorbed this week’s news with the kind of cognitive dissonance that markets do well. Palantir, whose AI platforms are central to battlefield intelligence and targeting workflows, posted Q4 revenues up 70% year on year and secured a $448 million defence contract.
Elbit Systems is up 61.6% year to date. The liability that Anthropic tried to build contractual walls against, and that the Pentagon insisted it had no intention of triggering, has now materialised in the most visible way possible. The question of whether AI companies can set limits on how their technology is used in warfare is no longer theoretical.
It is a live lawsuit, an active Pentagon investigation, and 69 unidentified children in a cemetery eight kilometres from a school called the Good Tree.
Keep up with Daily Euro Times for more updates!
Read also:
Dubai After Iran: Young Professionals Relocate for Speed, Not Sunshine
Sánchez Says No to War. Washington Says Pay for It Anyway.
Baalbek Clans and the Return of the State
