Google signs Pentagon AI deal

Google has signed a deal allowing the Pentagon to use its artificial intelligence models for classified work, according to reports published Tuesday. (reuters.com) (nytimes.com) The Information, cited by Reuters and The New York Times, said the agreement lets the Defense Department use Google AI for “any lawful government purpose” on classified networks. Classified systems handle some of the military’s most sensitive work, including mission planning and weapons targeting. (reuters.com) (nytimes.com) The deal surfaced one day after Google employees circulated a letter urging Chief Executive Sundar Pichai to refuse classified military AI work. Bloomberg reported that hundreds of employees signed it; other reports put the total at more than 600. (bloomberg.com) (cbsnews.com) In the letter, employees said AI systems can “centralize power” and make mistakes, and argued classified workloads would be harder for workers to monitor. They warned Google could become linked to surveillance, weapons systems, or other uses they would not be able to see from inside the company. (bloomberg.com) (cbsnews.com) This is not Google’s first Pentagon AI work. In December 2025, the Defense Department’s Chief Digital and Artificial Intelligence Office said Google Cloud’s Gemini for Government became the first frontier AI capability deployed on GenAI.mil, the department’s enterprise generative AI platform. (ai.mil) (googlecloudpresscorner.com) The Pentagon widened that push in July 2025, when the Chief Digital and Artificial Intelligence Office awarded Google, OpenAI, Anthropic and xAI contracts worth up to $200 million each for advanced AI work. The agency said the awards were meant to address “critical national security challenges.” (ai.mil) (cnbc.com) Google’s military posture had already shifted before this week. In February 2025, the company updated its AI principles and removed language pledging not to pursue weapons and surveillance applications. (cnbc.com) That change marked a break from the backlash Google faced in 2018 over Project Maven, a Pentagon program that used artificial intelligence to analyze drone footage. Employee protests then pushed the company to step back from that contract and publish tighter AI principles. (cbsnews.com) (cnbc.com) Google did not immediately restore the old red lines. Instead, the company is now moving deeper into the same classified market where OpenAI and xAI already have Pentagon arrangements, as the Defense Department builds out GenAI.mil for broad internal use. (reuters.com) (nytimes.com)