Google Expands Defense Partnerships with Classified AI Deal Amidst Internal Protests
Google has finalized a classified agreement with the U.S. Department of Defense, marking a significant expansion of the tech giant’s defense partnerships despite internal opposition. This agreement allows the Pentagon to utilize Google’s artificial intelligence models for classified military work.
The deal, signed on Monday, April 28, 2026, permits the Pentagon to use Google’s AI systems for “any lawful government purpose” on classified networks. These networks handle sensitive operations including mission planning and weapons targeting. Google now joins OpenAI and Elon Musk’s xAI in providing AI capabilities for classified defense work.
The agreement surfaced alongside a letter signed by over 600 Google employees. This includes researchers from the DeepMind AI lab, who urged CEO Sundar Pichai to reject the partnership. The employees expressed concern about potential uses in “inhumane or extremely harmful ways” once AI enters classified operations. Although the contract includes language stating AI should not be used for domestic mass surveillance or autonomous weaponry without human oversight, it also specifies that Google cannot veto lawful government operational decisions.
This marks a notable shift for Google, which withdrew from the Pentagon’s Project Maven program in 2018 following similar employee backlash over AI analysis of drone footage. In 2025, the Pentagon signed agreements worth up to $200 million each with major AI labs, including Google, OpenAI, and Anthropic. This was part of a broader effort to integrate advanced AI into defense operations.
Source: Bloomberg