600 Google employees sign joint protest: Refuse to let Gemini enter the Pentagon's classified military network

📄Full Article· Automatically extracted by trafilaturaGemini 翻譯1468 words

Over 600 Google employees, including more than 20 vice presidents and senior executives, have signed a joint letter to CEO Sundar Pichai, demanding that the company reject a new agreement to expand Gemini into the Pentagon's classified networks. Employees pointed out that no external entity can control AI systems within a classified environment, and existing safeguards are technically unenforceable. (Context: Y Combinator Startup Guide: What are the future development trends for AI Agents?) (Background: UC study on "AI brain fog": 14% of office workers driven crazy by Agents and automation, with a 40% higher turnover intention) Over 600 Google employees have signed a joint letter to CEO Sundar Pichai, with signatories including senior researchers from DeepMind, as well as over 20 vice presidents and senior executives, demanding that the company not authorize Gemini to enter the Pentagon's classified military networks. Let’s first summarize the cooperation between the two parties. At the end of 2022, Google, along with AWS, Microsoft, and Oracle, won a core cloud procurement project from the U.S. Department of Defense known as JWCC, with a total ceiling of $9 billion. In December 2025, the Gemini-based GenAI.mil platform was officially launched for use in unclassified environments. By March 2026, Gemini AI agents had been deployed to all 3 million personnel within the Department of Defense. All of the above were applications in the unclassified domain. However, the new agreement currently under negotiation aims to expand Gemini's capabilities into "classified" environments: air-gapped, closed networks physically isolated from external networks, specifically used for classified military operations. Simply put: Gemini is entering the combat command center. The disagreement at the negotiating table is clear. Google is attempting to draw red lines in the contract: prohibiting Gemini from being used to track domestic citizens or executing strike decisions without human intervention. The Pentagon's position, however, is "all lawful purposes"—a phrase that leaves no clear restricted zones and explicitly stipulates that external vendors shall not retain any control over their AI systems. The two conditions are in direct conflict. Google employees pointed out in the joint letter that the safeguards proposed by the company are "technically unenforceable." The initiators of the letter highlighted a fundamental issue: "Classified work, by definition, is opaque." This means that once Gemini enters a classified network, Google itself will not be able to see what it is doing. In an unclassified environment, Google can audit API calls, monitor model outputs, set guardrails, and intervene when problems are detected. In a classified environment, none of these exist. Employees listed specific concerns in the letter: profiling of individuals (using AI to build behavioral and identity models of target persons) and targeting innocent civilians. These use cases are not hypothetical but are existing patterns in AI-assisted military operations. The dilemma facing current Google management is that it cannot technically guarantee to employees that Gemini will not be used for specific purposes, because it simply cannot access that network to verify. In February 2025, Google quietly revised its AI Principles, removing clauses that explicitly prohibited the "development of weapons or surveillance AI." The reason given by DeepMind CEO Demis Hassabis was that a global competition for AI leadership is underway; Human Rights Watch and Amnesty International immediately condemned the revision. In the 2018 Project Maven case, 4,000 employees signed a petition and at least 12 resigned, eventually forcing Google to announce it would not renew the contract, which was then taken over by Palantir. The success of the Maven case was partly because AI military applications were still a fringe issue in 2018, and the cost to Google's brand outweighed the contract value. Eight years later, AI is a core competitive project for defense infrastructure, and Google, along with AWS and Microsoft, is competing for contracts across the board. Market position, political pressure, and the competitive landscape all point in the same direction: a single letter seems insufficient to counter the market interests behind such a massive contract.

Data Status✓ Full text extractedRead Original (動區 BlockTempo)

🔍Historical Similar Events· Keyword + Asset Matching6 items

2026-04-29

Google signs $200 million AI military contract with the Pentagon to implement Gemini; 600 employees protest in vain

Similarity 170%關鍵字 google/gemini同分類 zh

2026-04-23

Google admits partnership with Apple: Gemini to take over Siri features, rumored $1 billion annual payment to Apple

Similarity 170%關鍵字 google/gemini同分類 zh

2026-04-23

Google Maps integrates Gemini, launching three AI features focused on enterprise Agents

Similarity 170%關鍵字 google/gemini同分類 zh

2026-04-22