In 2018, roughly 4,000 Google employees signed a petition over the company’s involvement in Project Maven — a Pentagon program analyzing drone footage — and Google let the contract expire. The employee revolt worked. In 2026, more than 600 Google employees, including directors and vice presidents, published an open letter urging CEO Sundar Pichai to refuse a classified military AI deal. Google signed the deal within 24 hours. This time, the revolt failed. The difference between the two outcomes reveals something important about how power has shifted inside major technology companies.
What the Letter Said
The open letter, published April 27, argued that Google should follow Anthropic’s lead and refuse to allow its AI models to be used in classified military contexts without guardrails against domestic mass surveillance and autonomous weapons. The signatories — described as a cross-section of engineers, product managers, directors, and vice presidents — made a specific argument: “The only way to guarantee that Google does not become associated with such harms is to reject any classified workloads.”
The letter explicitly referenced Anthropic as a model for principled refusal. Anthropic had been designated a Pentagon supply chain risk for refusing the same terms Google was now being asked to accept. The employees argued that Google should be willing to accept similar consequences rather than compromise on AI ethics principles.
Why Google’s Leadership Held Firm
Google’s response was silence — followed by the deal. The company has not publicly acknowledged the employee letter, the contract, or offered any substantive response to the employees’ concerns. The closest it came to engagement was an internal memo about “responsible AI” and military partnerships that did not explicitly address the specific agreement employees were objecting to.
Labor researchers tracking the situation identified several structural reasons why the outcome differed from Project Maven:
- Google has internalized the Maven lesson — in the wrong direction. Following the 2018 revolt, Google systematically reduced the internal communication infrastructure that had enabled coordinated employee action: decommissioning internal mailing lists, removing the internal social network, and tightening controls on what could be discussed across teams. The employees who organized the letter in 2026 were operating without the same coordination tools that made the Maven petition effective.
- Tech layoffs have weakened employee leverage. Mass layoffs across the sector in 2024 and 2025 have made collective action harder. Workers who previously felt secure enough to publicly oppose leadership decisions now face a labor market where tech roles are less abundant and the consequences of internal dissent are more visible.
- AI is replacing engineers. As companies publicly acknowledge that AI can generate 65% or more of new code — as Snap disclosed in its April layoffs — the implicit threat underlying employee leverage changes. The cost of losing the engineers who sign protest letters is lower in a world where AI can absorb a significant share of their output.
- The financial stakes are much higher. The Project Maven contract was worth a few million dollars. The Pentagon’s classified AI infrastructure represents billions in potential revenue. The calculus for leadership has changed.
The Irony of Google’s Position
Google has committed up to $40 billion in investment in Anthropic — the company that was blacklisted by the same Pentagon that Google just signed a deal with. Google is simultaneously one of Anthropic’s largest financial backers and a competitor for the government contracts that Anthropic’s ethical stance cost it. That contradiction is unlikely to be resolved cleanly, and it will surface in any substantive discussion of Google’s AI ethics positioning going forward.
The classified deal itself includes language stating the AI should not be used for “domestic mass surveillance or autonomous weapons without appropriate human oversight” — language identical to what OpenAI included in its Pentagon contract. As noted previously, enforcement of these provisions on classified, air-gapped networks is structurally impossible from Google’s position: the company cannot monitor what queries are being run on systems it cannot access.
What Comes Next for Google Employees
Labor researchers quoted in Fortune’s coverage of the revolt note that the employees’ remaining leverage is limited but not zero. Regulatory scrutiny of military AI contracts is increasing. The EU AI Act’s provisions on high-risk AI systems are in full enforcement. Congressional oversight of classified AI deployments is expanding. External pressure, rather than internal organizing, may be the more viable avenue for accountability going forward.
Conclusion
The failure of Google’s employee revolt over the Pentagon AI deal is not just a story about one company’s governance. It’s a signal about a structural shift in how decisions about AI ethics get made in large technology companies — and who has the power to influence them. Browse our directory to follow Claude, ChatGPT, and the AI tools whose policy decisions are setting precedents for the entire industry.
