Google signed a classified agreement with the Department of Defense allowing Gemini to be used for "any lawful government purpose" on classified networks, becoming the third major AI company—after OpenAI and xAI—to do so, and filling the slot left when Anthropic was blacklisted for refusing equivalent terms. The contract requires Google to adjust its AI safety settings at the government's request and explicitly gives Google "no right to control or veto" how the DoD uses the models; non-binding language discourages autonomous weapons and domestic mass surveillance but is unenforceable. More than 600 Google employees, including some DeepMind executives, signed an open letter opposing the deal before it was signed, according to Bloomberg; Google did not respond to requests for comment.