Google Employees Urge CEO to Reject Pentagon’s Classified AI Deal

580 Google employees from DeepMind and Cloud signed an open letter to CEO Sundar Pichai urging the company to decline any contract that would let the U.S. Department of Defense deploy Gemini AI in classified settings. They cite lethal autonomous weapons and mass surveillance as unacceptable outcomes.

Business Insider

Google’s defense work began with Project Maven in 2018, which employee protests ended. After a pause, the firm resumed unclassified AI services in 2023 and supplied Gemini agents to the Pentagon’s three-million-person workforce in March 2024. Negotiations aim to expand use to “all lawful” classified tasks, echoing the halted Anthropic deal after the Pentagon labeled the startup a supply-chain risk.

East Bay Times

The letter lists over 20 directors, senior directors and vice presidents, plus dozens of senior engineers. About two-thirds are named, the rest anonymous. DeepMind researcher Sofia Liguori and Pentagon under-secretary Emil Michael, who noted plans to move Gemini to classified use, are among the signatories. Google already runs a non-classified AI contract on its genAI.mil platform.

East Bay Times

Google has not responded publicly, and the Pentagon declined comment. The Defense Department says commercial firms should not set usage limits during wartime, emphasizing “all lawful uses.” The Information reports negotiations for that language, while other sources say the Pentagon seeks to bar mass surveillance and autonomous weapons—provisions employees deem unenforceable. Neither side has confirmed terms, leaving the scope of any classified deployment uncertain.

East Bay Times

Employees say safeguards cannot guarantee AI will not be weaponized or used for surveillance, noting Department of Defense “specifically prohibits any controls.” They argue rejecting classified workloads is the only safeguard. Sofia Liguori said monitoring “air-gapped” systems is infeasible and described agentic AI as especially worrisome due to its high independence.

East Bay Times

• “We feel that our proximity to this technology creates a responsibility to highlight and prevent its most unethical and dangerous uses.” — Google AI employees, open letter

Gizmodo

• “As people working on AI, we know that these systems can centralize power and that they do make mistakes.” — Google AI employees, open letter

East Bay Times

• “We want to see AI benefit humanity; not to see it being used in inhumane or extremely harmful ways.” — Google AI employees, open letter

Gizmodo

• “Agentic AI is particularly concerning because of the level of independence it can get to.” — Sofia Liguori, DeepMind

East Bay Times

Negotiations are slated to continue through the summer. Employees say they will intensify outreach if Google moves forward with a classified agreement, and the decision could influence future corporate defense contracts and internal AI policy reviews across the tech sector.

Business Insider