The US strike on a primary school in Iran’s Minab city that killed 157 people, including 123 children, was allegedly linked to a series of intelligence failures, outdated satellite imagery and overdependence on artificial intelligence systems, according to a Bloomberg investigation.
The attack on Shajarah Tayyebeh Primary School took place on February 28, the first day of the US-Israeli strikes against Iran. Two cruise missiles hit the school compound, killing several students and civilians.
The United Nations has said there are reasonable grounds to believe that the attack may have amounted to a war crime involving the targeting of a civilian object.
Why Was A School Targeted?
The school’s location appears to have been at the centre of the intelligence failure.
According to the investigation, the land where the school stood had earlier been part of an Islamic Revolutionary Guard Corps (IRGC) military compound. However, the area had reportedly changed significantly before the strike and was operating as a school.
Satellite images showed that by 2017, walls and separate entrances had been built to separate the school from the nearby military facility. Images from 2018 showed a football field, colourful walls and markings linked to children’s activities.
Despite these visible changes, the location continued to be listed as an IRGC facility in the main military intelligence database used for selecting targets.
Bloomberg’s report, based on interviews with more than two dozen current and former US officials familiar with the Pentagon investigation, described the incident as a “cascade of preventable failures”.
Old Data Remained In Targeting System
The report said US intelligence analysts had noticed changes at the site as early as 2019. However, information about those changes was stored in a separate database that was not connected to the primary intelligence system used during the targeting process.
Because of this, officials involved in the Pentagon inquiry found that outdated information continued to influence the target assessment.
The satellite imagery used during the process also reportedly did not reflect the school’s transformation, meaning the data available to those making decisions did not match the actual situation on the ground.
AI System Added To Concerns Over Targeting Process
The investigation also looked at the Pentagon’s Maven Smart System, an artificial intelligence-enabled platform developed by Palantir Technologies.
The system is designed to analyse large amounts of intelligence and help military personnel identify patterns and possible targets. According to the report, it can process more than 150 different data inputs.
However, officials cited by Bloomberg said the concern was not that AI itself selected the school as a target.
Instead, some personnel at the US Central Command reportedly placed too much trust in the system and expected it to identify outdated information or inconsistencies in the intelligence being used.
Palantir rejected suggestions that its software was responsible for the strike. A company spokesperson told Bloomberg that Palantir was not responsible for the underlying data or identifying intelligence shortcomings, adding that there was no evidence its software was at fault.
Pressure For Rapid Military Action
The report also pointed to the pace of the military campaign as another factor.
The Trump administration’s push for a major aerial assault reportedly reduced the time available for finalising and reviewing targets. More than 1,000 Iranian targets were reportedly struck during the first 24 hours of the campaign.
Investigators are examining whether outdated intelligence, reduced civilian-protection staffing and dependence on automated systems combined to create a breakdown in the targeting process.
You May Also Like To Watch: