A report indicates that Google has joined a technology alliance making agreements with the U.S. Department of Defense, allowing AI models to be used in secret security missions encompassing "all lawful purposes" amid scrutiny over ethics and weapons control.
The Information news agency, citing insider sources, reported that Google, under parent company Alphabet, has reached a secret agreement with the U.S. Department of Defense permitting government agencies to use artificial intelligence (AI) models within classified networks. This marks a significant step placing Google alongside OpenAI and Elon Musk's xAI, which previously made similar agreements.
The agreement states that the Pentagon can use Google's AI for "any lawful government purpose." These classified networks handle highly sensitive tasks ranging from mission planning to weapons targeting.
Earlier in 2025, the Pentagon signed agreements worth up to $200 million with leading AI labs including Anthropic, OpenAI, and Google, aiming to have these companies deploy AI systems on classified networks without the standard restrictions applied to general users.
Google's contract includes a key condition requiring the company to assist in adjusting AI security settings and content filters as requested by the government. However, it explicitly states that "the parties agree the AI system is not intended, nor should it be used, for domestic mass surveillance or as autonomous weapons (including target selection) without appropriate human oversight and control." Conversely, the contract also specifies that Google has no authority to control or veto lawful government operational decisions.
A Google spokesperson confirmed that the company supports government agencies in both classified and general projects, maintaining a consensus that AI should not be used for surveillance or uncontrolled autonomous weapons. They view granting API access to commercial AI models through Google's infrastructure as a responsible approach to national security.
Meanwhile, the U.S. Department of Defense, renamed the "Department of War" by former President Donald Trump, declined to comment on this matter but has previously stated it has no interest in using AI for spying on Americans or creating autonomous weapons without human involvement. The department aims to use AI in "all lawful forms."
Google's move follows conflicts earlier this year when startup Anthropic, creator of Claude, refused to unlock safeguards preventing AI use in autonomous weapons or surveillance, leading the Pentagon to label the company a supply chain risk at that time.