The US military is exploring the use of generative AI chatbots to assist in targeting decisions, according to a Defense official speaking to MIT Technology Review. Under this approach, classified AI systems would analyze lists of potential targets and rank them based on factors like aircraft positioning, with human operators then verifying the recommendations. This represents a shift from traditional AI systems like Maven, which has used computer vision to identify targets from drone footage since 2017.

OpenAI's ChatGPT and xAI's Grok are among the models being considered for classified military use following recent agreements between these companies and the Pentagon. The official described this as a potential use case rather than current practice, though the military has been rapidly expanding AI deployment across operations. Anthropic's Claude has already been integrated into existing military systems and reportedly used in operations in Iran and Venezuela before being designated a supply chain risk by the Defense Department.

While generative AI could accelerate the targeting process by providing conversational interfaces to analyze data, it introduces new challenges around verification. Unlike Maven's visual interface that required direct data inspection, AI chatbot outputs are easier to access but harder to validate. The military's AI expansion comes amid heightened scrutiny following a recent strike on an Iranian school that killed over 100 children, with investigations ongoing into what role, if any, AI systems played in that incident.