Key Takeaways
- Over 600 Google employees, including 20+ directors and VPs, signed a letter opposing classified military AI work
- Google is negotiating with the Pentagon to extend Gemini AI into classified domains beyond its existing non-classified contract
- Employees argue proposed safeguards against misuse are technically unenforceable under Pentagon policy
More than 600 Google employees signed a letter Monday demanding the company reject a proposed Pentagon deal that would deploy its Gemini AI model in classified military operations. The signatories include workers from Google DeepMind, Cloud, and other divisions. Over 20 directors, senior directors, and vice presidents added their names.
The letter, addressed to CEO Sundar Pichai, arrives as Google actively negotiates with the U.S. Department of Defense. The company already holds a contract for non-classified workloads through a program called genAI.mil. The proposed new deal would extend Gemini into classified settings.
The Core Dispute: Opacity and Accountability
Employees argue that classified workloads create an accountability black hole. By definition, classified operations shield their AI applications from public scrutiny.
“Classified workloads are by definition opaque. Right now, there's no way to ensure that our tools wouldn't be leveraged to cause terrible harms or erode civil liberties away from public scrutiny. We're talking about things like profiling individuals or targeting innocent civilians.”
— Anonymous Google employee organizing the letter
According to the letter organizers, Google has proposed contract language that would prevent Gemini from being used for domestic mass surveillance or autonomous weapons without human control. The Pentagon, however, wants broad "all lawful uses" wording. The military argues this flexibility is necessary for operational purposes.
Employees say Google's proposed safeguards are technically unenforceable. They point to Pentagon policy that prohibits outside entities from imposing controls on its AI systems. Once Gemini enters classified domains, Google loses oversight.
“If leadership is truly serious about preventing downstream harms, they must reject classified workloads entirely for now.”
— Second letter organizer
Filling the Anthropic Void
Google is one of several companies competing to become the Pentagon's go-to AI provider. The opportunity opened after AI startup Anthropic fell out of favor with the Defense Department.
Anthropic has sued the Pentagon over its designation as a "supply-chain risk." That designation came after Anthropic requested that its technology not be used for mass surveillance in the United States or for automated warfare. The company's ethical stance cost it Pentagon access. Google now sees a business opportunity.
Related: Government tech security and national defense implications
Project Maven's Ghost
This isn't Google's first internal revolt over military AI. The current campaign draws direct inspiration from a 2018 employee movement that successfully killed Project Maven. That Pentagon program aimed to integrate AI into drone operations.
The 2018 protest worked. Thousands of employees signed a petition, and dozens resigned. Google ultimately chose not to renew the Project Maven contract. The company also published AI principles that included a pledge not to develop AI for weapons.
Seven years later, the stakes have shifted. AI capabilities have advanced dramatically. Government AI spending has ballooned. And Google faces intense competition from Microsoft, Amazon, and others for lucrative defense contracts.
What Happens Next
Google has not publicly responded to the employee letter. The company's decision will signal how much weight internal dissent carries in 2025 versus 2018. The tech labor market has cooled. Mass layoffs have reshaped power dynamics between workers and executives.
The Pentagon contract negotiations continue. Google must choose between a lucrative classified AI deal and its workforce's moral objections. The 2018 playbook showed employee pressure can work. Whether that remains true in today's environment is an open question.
Related: Major AI companies expanding into new markets and partnerships
Logicity's Take
Frequently Asked Questions
What is Google's current contract with the Pentagon?
Google holds a contract for non-classified workloads through a program called genAI.mil. The proposed new deal would extend Gemini AI capabilities into classified military operations.
Why did Anthropic lose its Pentagon contract?
Anthropic requested that its AI not be used for mass surveillance in the U.S. or automated warfare. The Pentagon designated it a "supply-chain risk," and Anthropic has since sued the department over that designation.
What was Project Maven?
Project Maven was a 2018 Pentagon program to integrate AI into drone operations. Google employee protests, including thousands of petition signatures and dozens of resignations, led Google to abandon the contract.
What safeguards has Google proposed for the Gemini contract?
Google proposed contract language preventing Gemini from being used for domestic mass surveillance or autonomous weapons without human control. The Pentagon wants broader "all lawful uses" language instead.
Why do employees say Google's safeguards won't work?
Pentagon policy prohibits outside entities from imposing controls on its AI systems. Once Gemini enters classified domains, employees argue Google loses any ability to enforce its proposed restrictions.
Need Help Implementing This?
Source: Tech-Economic Times / ET
Huma Shazia
Senior AI & Tech Writer
Produced with AI assistance and reviewed by the Logicity editorial team. Learn more in our Editorial Policy.
