WASHINGTON (Realist English). US Central Command, responsible for combat operations in the Middle East, has made multiple changes to its target selection process and expanded the use of artificial intelligence following a deadly strike on a school in the Iranian city of Minab. Bloomberg reported this, citing US officials and sources familiar with the situation.
Essence of the Changes
The previously unreported changes at CENTCOM focus on three key areas:
- Refining processes and operating procedures for updating and verifying planned targets
- Adding new open-source data to better track the movement of civilians
- Dozens of updates to the Maven Smart System — an AI-enabled battle management programme developed by defence contractor Palantir Technologies
According to sources, the new capabilities include ways for the Maven system to sort intelligence to determine a building’s function. Cameron Stanley, head of the Pentagon’s Digital and AI Office, confirmed in an interview that the updates include the addition of open sources helping track civilian maritime traffic, which has become more important as the conflict has evolved.
Investigation: What the Pentagon Found
The tragedy occurred on 28 February 2026, when two Tomahawk missiles struck the Shajarah Tayyebeh primary school in Minab, killing more than 150 people, including at least 123 children. The school was located on territory that had previously been part of a military complex, but was separated by walls and functioned as an educational institution.
On the day of the strike, the US hit 1,000 targets in a surprise attack. According to Stanley, such massive firepower was made possible by the Maven Smart System.
The Pentagon’s investigation revealed a series of preventable errors: outdated intelligence, shortened target verification timelines, reduced civilian protection personnel, and excessive reliance on AI systems. Some CENTCOM staff expected Maven to identify outdated or contradictory information, but it did not.
Changes in Doctrine and the Maven System
After the strike, Palantir added new capabilities to Maven, including AI agents that re-check intelligence and have already identified new anomalies. One official noted that the new tools allow “reviewing source intelligence to identify factors disqualifying a target and flagging inconsistencies that human analysis might have missed.”
In April 2026, the Pentagon also quietly revised its target selection doctrine, opening the way for broader AI participation in critical decision-making. The new edition, in particular, acknowledges the need to “exercise caution to avoid excessive reliance on AI outputs without human oversight,” while simultaneously stating that algorithms can analyse intelligence “at scales and speeds far exceeding human capabilities.”
Despite the changes, some defence officials and experts warn that the probability of a repeat of Minab remains too high. The full investigation report, launched in March, has still not been published, although sources say it has effectively been completed for several months.
International Reaction
The UN-supported fact-finding mission stated on 17 September that there are reasonable grounds to believe the US committed war crimes during two airstrikes in Iran in February, including the strike on the school in Minab. The mission qualified the strike as an indiscriminate attack that resulted in civilian casualties.
The Pentagon declined to comment on the UN report, saying the investigation is ongoing. White House spokesperson Anna Kelly, commenting on the report, said the UN Human Rights Council “has done nothing for human rights,” adding that Iran is the only party to the conflict that has “committed war crimes.”
