According to the Post, the letter says that “The only way to guarantee that Google does not become associated with such harms is to reject any classified workloads. Otherwise, such uses may occur without our knowledge or the power to stop them.” Anthropic is currently in a legal battle with the Pentagon over being designated a “supply chain risk,” after refusing to loosen guardrails around how the US military can use its AI models, with support from across the tech industry, including employees at Google.

The letter specifically references a recent report by The Information that said Google and the Pentagon are discussing a deal for deploying its Gemini AI in classified settings. Microsoft already has deals to provide AI services in classified environments, and OpenAI announced a renegotiated agreement with the Pentagon in February.