Workers from DeepMind and Cloud signed a letter citing concerns about lethal autonomous weapons and mass surveillance
More than 600 Google $GOOGL employees sent a letter to CEO Sundar Pichai this week demanding the company refuse all classified AI work with the Pentagon, citing concerns that such arrangements would prevent workers from knowing how the technology is used.
"We want to see AI benefit humanity; not to see it being used in inhumane or extremely harmful ways," the employees wrote. "This includes lethal autonomous weapons and mass surveillance but extends beyond."
Signatories came from Google's DeepMind and Cloud divisions. Their letter pointed to reporting by the Information that Pentagon officials had been discussing with Google the possibility of deploying its Gemini model in classified environments — an arrangement resembling the contract OpenAI struck with the Defense Department in February.
"The only way to guarantee that Google does not become associated with such harms is to reject any classified workloads," the letter said. "Otherwise, such uses may occur without our knowledge or the power to stop them."
A request for comment sent to Google went unanswered, Business Insider reported.
Among those who signed, Decrypt noted, are over 18 senior-level staff whose titles span from principal to vice president; the letter also pushes for full disclosure on current Defense Department contracts and for Google to create a standing ethics board — one that includes employees — to evaluate any proposed military deals going forward.
The letter arrives at a fraught moment for the AI industry's relationship with the Defense Department. After Anthropic attempted to insert contract language prohibiting its technology from being applied to mass surveillance or fully autonomous weapons, the Pentagon severed ties with the company in February. Whether that decision was lawful is now being contested in federal court, according to The Washington Post.
Google has faced similar internal pressure before. Tech companies including Google have been rapidly expanding their military partnerships since President Donald Trump's election victory, abandoning years of policies that restricted defense work. When thousands of Google workers petitioned against Project Maven back in 2018, the company ultimately walked away from the Defense Department program, which had been applying AI to the interpretation of drone imagery. Following that withdrawal, Google codified a set of AI principles that explicitly ruled out applications involving weapons systems or surveillance.
Those guardrails were quietly stripped from Google's AI principles the following year, and the company separately reached an agreement allowing the Pentagon to make use of its Gemini AI. By March, Google had committed to furnishing Defense Department operations with AI agents, though in an unclassified context; staff at DeepMind had already been told in a January internal meeting that similar agreements were coming.
Senior Defense Department officials have said that the military should be allowed to use commercial AI in any situation that is legal. They argue this approach keeps their options open while staying within U.S. law and military rules.
