Alphabet’s Google signed a classified artificial intelligence agreement with the United States Department of Defense on 28 April 2026. The arrangement allows the Pentagon to use Google’s AI models for classified work and for any lawful government purpose.
Contract Framework
The deal is an amendment to Google’s existing contract with the Department of Defense. The Pentagon had signed agreements worth up to $200 million each with Anthropic, OpenAI and Google in 2025 for AI-related work. The contract includes a clause stating that the AI system is not intended for domestic mass surveillance or autonomous weapons, including target selection, without appropriate human oversight and control. The same language states that the agreement does not give any right to control or veto lawful government operational decision-making.
Google Models In Government Use
Google’s Gemini for Government 3.1 Pro and 3.0 Flash models were added on 27 April 2026 to the U.S. Department of War’s GenAI.mil platform. GenAI.mil is a secure artificial intelligence platform used by the U.S. defence establishment.
- Gemini for Government 3.1 Pro is a Google model used for government applications.
- Gemini for Government 3.0 Flash is a Google model used for faster AI processing tasks.
- The models are linked with research, vision and reasoning functions.
- The platform is designed to reduce hallucinations in AI outputs.
Employee Response And Earlier Context
More than 580 Google employees, including staff from DeepMind and Cloud, sent an open letter to CEO Sundar Pichai on 27–28 April 2026. The letter asked Google to refuse classified Department of Defense AI work. Google ended its Project Maven contract in 2018 after internal employee protests. Project Maven was a Pentagon programme that used AI to analyse drone footage.
Last Modified: April 28, 2026