Google Employees Raise Alarms Over Military AI Use

Concerns over military use of AI are stirring tensions in Silicon Valley, as employees at major tech firms question how their work could be leveraged by the US government.

According to The New York Times, over 100 Google employees working on AI recently urged the company to set strict limits on how its systems might be used in defense applications.

They wrote to Jeff Dean, chief scientist at Google DeepMind, raising alarms about Gemini AI being deployed for mass surveillance of US citizens or for autonomous weapons without human oversight.

The debate mirrors a similar standoff at Anthropic, an American AI firm whose models are used in classified military programs.

The US Department of Defense reportedly holds a $200-million contract with the company and has pushed for broader access to its AI models.

Anthropic has resisted, seeking guarantees that its technology will not be used for domestic mass surveillance or fully autonomous weapons.

Reports indicate the Pentagon may invoke the Defense Production Act, a law allowing the government to require companies to prioritize national defense needs, if Anthropic does not comply.

U.S. Army Sgt. Michael Morales, assigned to 41st Field Artillery Brigade, updates Soldier information in a battalion aid station during Saber Guardian 25, Cincu Training Area, Romania, June 13, 2025. The aid station enables rapid triage, treatment, and evacuation of casualties, ensuring readiness and lifesaving support in austere operational environments. (U.S. Army photo by Spc. Hunter Carpenter) Demonstrating global deterrence and the U.S. Army’s ability to rapidly deploy U.S.-based combat power in Europe and the Arctic region alongside Allies and partners, DEFENDER 25 brings U.S. troops together with forces from 29 Allied and partner nations to build readiness through large-scale combat training from May 11-June 24, 2025. DEFENDER 25 increases the lethality of the NATO alliance through large-scale tactical training maneuvers and long-range fires, builds unit readiness in a complex joint, multinational environment and leverages host nation capabilities to increase the U.S. Army’s operational reach. During three large-scale combat training exercises—Swift Response, Immediate Response, and Saber Guardian—Ally and partner forces integrate and expand multi-domain operations capability, demonstrating combined command and control structures and readiness to respond to crisis and conflict. — A personnel updates Soldier information in a deployed battalion station during an exercise. Photo: Spc. Hunter Carpenter/US Army

Google’s Defense AI Programs

Google maintains multiple ongoing projects with Pentagon and US defense firms focused on AI and cloud infrastructure.

In December 2025, the US Department of Defense launched GenAI.mil, a platform offering generative AI tools to nearly 3 million military personnel, civilian staff, and contractors.

The system integrates a version of Gemini tailored for government use, helping with document drafting, policy summaries, risk analysis, and media review.

Google’s work with the defense sector has grown through additional partnerships.

In November 2025, General Dynamics Information Technology teamed up with Google Public Sector to provide AI-enabled cloud infrastructure for US defense and intelligence agencies.

Earlier that year, Lockheed Martin announced plans to integrate Google Cloud’s Vertex AI into its AI Factory program, accelerating the development of generative AI models for national security, aerospace, and scientific applications.