Over 560 Google workers signed a letter pleading with their CEO to say no. <\/p>\n
He said yes anyway.<\/p>\n
The Deal Nobody Was Supposed to See<\/p>\n
Google has signed a classified agreement with the U.S. Department of Defense that lets the Pentagon use its AI models for \u201cany lawful government purpose.\u201d <\/p>\n
The Information<\/a> broke the story on Tuesday, April 28, 2026, citing a single source familiar with the deal.<\/p>\n The timing couldn\u2019t be more awkward. <\/p>\n Just one day earlier, more than 560 Google employees sent an open letter to CEO Sundar Pichai<\/a> urging him to block the Pentagon from accessing Google\u2019s AI for classified work. They warned the technology could be used in \u201cinhumane or extremely harmful ways.\u201d<\/p>\n Pichai clearly had other plans.<\/p>\n What the Contract Actually Says<\/p>\n According to The Information\u2019s report<\/a>, the deal includes language stating that Google\u2019s AI shouldn\u2019t be used for domestic mass surveillance or autonomous weapons without \u201cappropriate human oversight and control.\u201d <\/p>\n That sounds reassuring on paper.<\/p>\n But here\u2019s the catch. The contract also states that Google has no \u201cright to control or veto lawful government operational decision-making.\u201d <\/p>\n In other words, those restrictions aren\u2019t enforceable. Google can\u2019t actually stop the Pentagon from doing anything it considers legal.<\/p>\n The deal also requires Google to help adjust its AI safety settings and filters whenever the government asks. A Google spokesperson told Reuters<\/a> the agreement is an amendment to an existing government contract. The company called it a \u201cresponsible approach to supporting national security.\u201d<\/p>\n Google Joins a Growing Club<\/p>\n This deal puts Google alongside OpenAI<\/a> and Elon Musk\u2019s xAI<\/a>, both of which have their own classified AI agreements with the Pentagon. <\/p>\n According to Reuters<\/a>, the Pentagon signed deals worth up to $200 million each with major AI labs in 2025.<\/p>\n Classified networks handle some of the military\u2019s most sensitive work. T<\/p>\n hat includes mission planning and weapons targeting. The Pentagon has said it doesn\u2019t intend to use AI for mass surveillance of Americans or for weapons without human involvement. <\/p>\n But it wants full flexibility to use AI for any lawful purpose, and it\u2019s not willing to let tech companies draw those lines.<\/p>\n What Happened to Anthropic <\/p>\n One major AI company tried to push back. <\/p>\n Anthropic, the maker of Claude, insisted on keeping guardrails<\/a> that prevented its models from being used for fully autonomous weapons or domestic mass surveillance.<\/p>\n The Pentagon didn\u2019t like that. <\/p>\n