The US military is testing five large language models (LLMs) of generative AI to see if they can be used to plan responses to potential global conflicts or provide faster access to internal information. The tests are being conducted by the digital and AI office at the Pentagon and involve giving the LLMs classified operational information to generate responses on real-world matters.

One test saw an AI model deliver a request for information in just 10 minutes, a significant improvement on the days it can take with multiple personnel. The tests are set to run until July 26 and are part of the military's ongoing study of AI's potential capabilities in warfare.