Pentagon's AI Targeting System Under Scrutiny After Tragic Iranian School Strike

On the first day of the U.S.-Iran war, a Tomahawk cruise missile struck Shajareh Tayyebeh elementary school in Minab, southern Iran, killing at least 168 people, including over 100 children. The incident has cast a shadow over the Pentagon’s use of artificial intelligence in targeting systems, sparking urgent inquiries from lawmakers and raising questions about the reliability of AI in military operations.

Questions Arise Over AI's Role in Targeting

More than 120 House Democrats have sent a letter to the Pentagon, demanding clarity on whether the Maven Smart System, an AI-driven targeting platform, was used to identify the school as a target. The system, built by Palantir Technologies under a $1.3 billion contract, is designed to process vast amounts of data, including satellite imagery and drone feeds, to generate strike coordinates in near real-time.

“Was artificial intelligence, including the use of the Maven Smart System, used to identify the Shajareh Tayyebeh school as a target?” the lawmakers ask. They also want to know if a human verified the accuracy of the target before the strike.

Outdated Intelligence Blamed for Tragedy

According to sources briefed on preliminary findings, U.S. Central Command created targeting coordinates using outdated intelligence provided by the Defense Intelligence Agency. This intelligence did not reflect the school’s presence, which had been established several years prior and was active on social media and had its own website.

Ukrainian drone operators, who have experience with semi-autonomous targeting systems, suggest that the failure likely stems from stale human-curated data rather than an AI malfunction. Ihor Matviyuk, director of Aero Center, a Ukrainian drone company, says, “The main problem was not the AI — it was how close the military object was to the school.”

Industry Context and Implications

The incident highlights the growing reliance on AI in modern warfare and the critical need for accurate, up-to-date intelligence. Maven’s ability to compress kill-chain reasoning and decision-making into the fastest timelines ever seen on the battlefield comes with significant risks, especially when human oversight is lacking.

Matviyuk’s experience in Ukraine underscores the limitations of AI in identifying camouflaged targets. “Automatic targeting allows us to capture less than half of the targets, not more,” he notes. This aligns with the Defense Department’s own data, which shows that Maven can correctly identify objects at roughly 60% accuracy.

As the investigation continues, the Pentagon faces pressure to provide a clear explanation and to reassess the role of AI in military operations. The tragic strike in Minab serves as a stark reminder of the potential consequences of relying too heavily on technology without adequate human verification.

References

• Deadly Iran school strike casts shadow over Pentagon’s AI targeting push