A 20-year-old Iraqi student, Abdul-Rahman al-Rawi, was killed by a US missile strike in al-Qaim, Iraq, in February 2024, in what a joint investigation suggests is the first acknowledged civilian death from an AI-assisted airstrike. The attack was part of a series of 85 coordinated strikes against Iranian-backed militias, which US officials initially hailed as a success due to the use of advanced AI targeting technology.
Abdul-Rahman's brother, Anmar, described the grisly aftermath: 'It took us two days to gather all of my brother’s remains.' The US later sent a letter of condolence, accepting that Abdul-Rahman was killed by mistake. However, US Central Command (Centcom) stated it had 'no way of knowing' whether AI was used in that specific strike, a claim experts say raises serious concerns.
Professor Jessica Dorsey of Utrecht University warned that the inability to trace the strike's targeting process implies a lack of record-keeping, while Dr Elke Schwarz of Queen Mary University noted that AI systems can outpace human judgement. The strikes, which also wounded up to 15 people and allegedly killed medical personnel, were in retaliation for a drone attack on a US base in Jordan.
The AI system used, Project Maven, employs computer vision algorithms to identify targets from satellite imagery. Centcom’s former chief technology officer, Schuyler Moore, confirmed in interviews that the February strikes were aided by machine-learning algorithms, marking the first public admission of AI use in individual US strikes. The incident has intensified debate over the role of AI in military operations and civilian protection.
