Translation result
As Google negotiates with the Department of Defense over using its Gemini AI tool in classified environments, some 600 employees sent an open letter to CEO Sundar Pichai urging him to halt the deal. They argue that because AI can concentrate power and produce errors, it should not be deployed on classified work that could enable military or surveillance uses.
The letter was signed largely by employees who work directly on Google’s AI systems. They voiced concern about the ongoing talks with the U.S. Department of Defense and stressed that those closest to the technology have a duty to warn about its riskiest applications. In particular, they flagged contract language that would allow the DoD to use the Gemini model in classified settings “for all lawful purposes.”
Employees say that specific phrasing is the main worry. While it appears legally bounded on its face, it can be interpreted to cover surveillance, tracking, intelligence analysis, and operational support. They argue that such broad language cannot prevent misuse and that the only reliable solution is to refuse classified work altogether.
Similar disputes have emerged at other AI firms. During negotiations over a contract worth approximately 294.5 billion KRW (about $220.9 million), Anthropic objected to language permitting its Claude model to be used for “all lawful purposes.” After talks broke down, the company entered legal disputes with the U.S. government. The DoD later reportedly designated Anthropic a supply-chain risk.
OpenAI also revised contract language with the Department. Reports say the amended clause bars using AI for deliberate tracking, surveillance, or monitoring of U.S. citizens or nationals, and restricts surveillance that relies on commercially acquired personal or identifiable information. The change reflects industry recognition that, as military uses of AI expand, at least minimal guardrails are necessary.
Google’s internal backlash echoes a controversy from 2018. Back then, employees protested participation in a military project, and Google subsequently committed in its AI principles not to pursue development likely to cause harm. The company pledged not to design or deploy AI for weapons or surveillance, but critics say the wording and scope of those principles have shifted.
In the letter, employees warned that a wrong decision could inflict irreversible damage on Google’s reputation, business, and global role. They also pointed out that misuse of technology has already resulted in loss of life and violations of civil liberties at home and abroad.
The debate goes beyond a labor dispute to a larger question about how far generative AI should be integrated with national security. How Google resolves its negotiations with the Department of Defense could shape future Big Tech AI ethics standards and government contracting practices.
TP AI Notice This article was summarized using a TokenPost.ai-based language model. Key details may be missing or inaccurate.
Most Commented