Gate News message, April 17 — Google is in talks with the U.S. Department of Defense to allow the Pentagon to use its Gemini AI models in classified settings, according to The Information. The proposed agreement would permit use of the models for all lawful purposes while restricting their deployment for domestic mass surveillance or autonomous weapons without appropriate human control.

Approximately 1.2 million Defense Department employees have already used Google's AI chatbot through GenAI.mil for unclassified tasks since December 1. The rollout includes "Agent Designer," enabling the Pentagon's roughly 3 million civilian and military personnel to build custom AI assistants without coding. Google is pricing "Gemini for Government" at 47 cents per agency annually under a General Services Administration deal, significantly undercutting comparable offers from OpenAI and Anthropic, which are each charging $1 per agency for one year.

The classified deal represents a strategic shift for Google. In 2018, the company declined to renew Project Maven, a Pentagon AI initiative analyzing drone footage, following employee protests. However, CEO Sundar Pichai has since signaled less tolerance for internal activism on government contracts. The Pentagon's push for Google partnership follows its designation of Anthropic as a "supply chain risk" over contract disputes concerning autonomous weapons and mass surveillance concerns, creating an opening for Google to expand its federal footprint.