Google has signed a classified agreement allowing the US Department of Defense to use its AI models for "any lawful government purpose," The Information reports. The deal places Google alongside OpenAI and xAI, which have also made classified AI agreements with the US government. The timing is awkward—employees demanded CEO Sundar Pichai block the Pentagon from using its AI just days earlier, citing concerns about "inhumane or extremely harmful" applications.

The agreement states Google's AI systems shouldn't be used for domestic mass surveillance or autonomous weapons "without appropriate human oversight." However, the contract explicitly doesn't give Google "any right to control or veto lawful government operational decision-making," suggesting the restrictions function more as informal commitments than binding legal obligations. The deal also requires Google to modify its AI safety filters when the government requests it. Anthropic faced a different outcome—blacklisted by the Pentagon for refusing to strip weapon and surveillance guardrails from its models.

A Google spokesperson said the company is "proud to be part of a broad consortium of leading AI labs and technology and cloud companies providing AI services and infrastructure in support of national security." This agreement amends Google's existing government deal. The arrangement underscores the growing tension between AI companies' safety pledges and military demands for fewer restrictions.