A new conflict over the military use of artificial intelligence has resurfaced inside Google. More than 560 employees sent an open letter to CEO Sundar Pichai urging him to refuse any use of Google's AI technology in classified U.S. military operations.
On April 27, Bloomberg reported that the staff emphasized in the letter that AI should serve humanity and must not be deployed in inhumane ways or used to cause serious harm.
They singled out lethal autonomous weapons and large-scale surveillance systems, arguing that the only way to eliminate those risks at their root is to refuse any use of AI in classified environments.
The move comes as clashes between the Pentagon and AI companies have intensified. Rival Anthropic pushed for guardrails to limit its models’ use in autonomous weapons and mass surveillance, but after disputes over those limits it faced pressure, including being designated a supply-chain risk.
Reports that Google was negotiating with the Pentagon to allow classified use of its Gemini model triggered the internal backlash. Employees warned that meaningful oversight and control of AI use are essentially impossible within classified systems.
The letter was led by DeepMind researchers; roughly 40% of signatories are from AI teams, with the remainder from Cloud and other groups.
Several senior leaders, including directors and vice presidents, also signed, signaling that concerns extend beyond rank-and-file employees and reflect organization-wide unease.
This isn't the first time Google staff have pushed back against military collaboration. In 2018 they protested the company's participation in the Pentagon’s drone-video analysis project, Project Maven, and Google ultimately chose not to renew the contract. The company later adopted AI principles pledging not to use its technology for weapons development or surveillance.
But Google has recently softened that stance, removing the line in its policy that committed it to avoid developing weapons or harmful technologies. DeepMind CEO Demis Hassabis said, "With the AI landscape changing rapidly, the role of tech companies in national security has also evolved."
Employees say the change could weaken the company’s ethical standards. One insider warned, "AI-driven mass surveillance can directly threaten civil liberties. This isn't a theoretical risk — it's already happening."
Reporter Chan Park cpark@aitimes.com
