Google has reportedly signed a classified AI agreement with the U.S. Pentagon, allowing its models to be used for sensitive government operations.
The deal places Google alongside OpenAI and xAI as approved providers for classified defense environments. According to reports, the Pentagon can use Google’s AI for lawful government purposes, while Google may also be required to adjust safety settings at government request.
The agreement reportedly bars domestic mass surveillance and autonomous weapons without human oversight, but preserves government control over operational decisions.
The move highlights growing Pentagon efforts to integrate frontier AI into defense, intelligence and mission planning systems.
