Full Breakdown
Google Employees Petition to Block Classified Pentagon AI Deal
4/27/2026, 10:43:15 PM
Core Event: Employees Urge Block of Classified Gemini Contract
Over 600 Google staff, including DeepMind and Cloud engineers, sent an open letter to CEO Sundar Pichai demanding that the company refuse a proposed Pentagon agreement to deploy its Gemini AI model in classified military operations. The petition argues that such work would lack oversight and could enable weaponized AI.
Background & Context: Prior Military AI Controversies
Google halted its 2018 Project Maven contract after employee protests and pledged not to develop AI for weapons, a stance later softened when the company removed explicit weapons language from its AI Principles. It currently supplies the Pentagon with non-classified AI via the genAI.mil program while negotiating the classified Gemini extension.
Key Figures & Groups: Internal Leaders of the Campaign
The letter was organized by DeepMind and Cloud engineers and signed by more than 20 senior leaders, including DeepMind chief scientist Jeff Dean. DeepMind co-founder Demis Hassabis publicly defended a broader defense role for AI firms. Jane Chung of Justice Speaks represents the signatories.
Data & Statistics: Scale of the Petition
More than 600 employees signed, with roughly two-fifths from the AI division and a comparable share from Cloud. The list includes over 20 directors, senior directors or vice presidents and dozens of unnamed staff across other units.
Official Statements & Responses
Google has not yet responded to the letter, according to Jane Chung. The company confirmed a non-classified Pentagon contract via genAI.mil but declined comment on the classified Gemini proposal. Google ended Project Maven in 2018 and, last year, removed the weapons-and-surveillance clause from its AI Principles. Hassabis said the “technological landscape has fundamentally changed,” implying a duty to support national defense.
Criticism & Opposition
Employees contend that classified workloads are “by definition opaque,” limiting oversight and raising the risk of lethal autonomous weapons. They cite Anthropic’s legal dispute with the Pentagon over “supply-chain risk” designations as evidence that industry safeguards are insufficient. The letter warns that harmful uses could occur “without our knowledge or the power to stop them.”
Verbatim Quotes
- “As people working on AI, we know that these systems can centralize power and that they do make mistakes,” — Google employees, letter
- “According to the Post, the letter says that “The only way to guarantee that Google does not become associated with such harms is to reject any classified workloads.” — Google employees, letter
- “We want to see AI benefit humanity; not to see it being used in inhumane or extremely harmful ways,” — Google employees, letter
- “Gemini, the Military and a Line in the Sand DeepMind’s chief scientist Jeff Dean has emerged as the most prominent internal voice on the issue, publicly stating that mass surveillance violates constitutional protections and carries a chilling effect on free expression.” — Jeff Dean, internal comment
Conflicting Reports & Gaps
Sources report differing signatory counts—“more than 600,” “over 560,” and “600+.” The terms of the proposed classified Gemini contract, including its scope, duration and oversight provisions, have not been publicly disclosed.
What's Next
CEO Sundar Pichai must decide whether to approve the classified Gemini extension. A rejection would meet employee demands and could reshape Google’s defense business; approval may spur further internal dissent and intensify industry scrutiny of AI-military partnerships.
