Google Signs Classified Pentagon Deal for ‘Any Lawful’ Use of AI

Google has reportedly signed a classified deal with the US Department of Defense (DoD) that permits the military to use its AI models for “any lawful government purpose,” according to The Information. The agreement comes a day after Google employees demanded CEO Sundar Pichai block Pentagon AI use over concerns of inhumane applications.

Key Terms of the Deal

• No veto power: The contract explicitly states Google does not have “any right to control or veto lawful government operational decision-making,” meaning restrictions on weaponization and surveillance are effectively non-binding promises, not legal obligations.
• Adjustments at request: Google must assist in modifying its AI safety settings and filters at the government’s request.
• Scope: AI systems should not be used for domestic mass surveillance or autonomous weapons “without appropriate human oversight and control” — but enforcement relies on mutual good faith.

Context

Google joins OpenAI and xAI in making classified AI deals with the US government. Anthropic was blacklisted by the Pentagon earlier for refusing to remove weapon and surveillance guardrails. A Google spokesperson stated: “We are 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.”

Implications for Developers

Developers using Google’s AI models (e.g., through Vertex AI or Cloud AI) should be aware that safety filters can be overridden for government customers. This may affect trust in Google’s AI governance, especially for those relying on consistent safety guardrails. Expect increased scrutiny on how AI agents are used in defense contexts.