CENTCOM Overhauls AI Targeting Protocols After Deadly School Strike
The US Central Command has rewritten its protocols for using artificial intelligence to select and vet lethal targets in combat following a deadly school strike that killed approximately 156 people, including 120 children, at an elementary school in Iran in February.
The Minab school strike on February 28 was the single deadliest civilian incident during Operation Epic Fury, the US military campaign against Iran. Seven months later, CENTCOM announced changes to its AI targeting protocols, focusing on three areas: vetting targets before a strike is authorized, tracking civilian presence near potential targets in real-time, and updating the underlying AI systems' data.
The overhaul centers around Palantir Technologies' Maven Smart System, the AI platform relied upon by the Pentagon to accelerate the targeting cycle. The system received extensive upgrades, including new capabilities to accurately classify building functions. CENTCOM also deployed AI agents to continuously re-evaluate intelligence, rather than relying on static snapshots that may be days or weeks old.
The updated Maven Smart System enabled the military to strike around 1,000 targets in a single day, a pace unthinkable in previous conflicts. However, Congressional scrutiny has intensified since the strike, with Senate officials raising pointed questions about how AI-generated targeting data is managed and stored.