A Russian drone powered by an autonomous AI system killed three Ukrainian civilians in early July in the Zaporizhzhia region, in what experts believe represent the first deaths inflicted by a fully autonomous AI-guided weapon with no human in the loop, according to a New York Times investigation published August 24. The attack marked a dangerous threshold in the use of artificial intelligence in warfare. Unlike previous AI-assisted strikes where humans planned and approved targets, this drone was programmed with target parameters and left to independently identify and strike on its own. Ukrainian officials confirmed the drone was equipped with an Nvidia Jetson Orin minicomputer -- a chip not officially sold in Russia but available through resale markets. 'This is a risk for the whole world,' a Ukrainian commander told the Times. 'In a few years, we will be living in a Terminator movie.' Analysts said the incident crosses a critical line by leaving machines to interpret the laws of war and determine what constitutes a legitimate target. Experts noted that while AI has been used in the final stages of human-planned strikes before, this appears to be the first confirmed case where the entire decision-making chain -- from target identification to lethal action -- was performed by AI without human authorization.

Sources