Google approves Pentagon to use AI tools in classified settings

Google has authorized the U.S. War Department to use its artificial-intelligence tools in classified environments.

The agreement accelerates the Pentagon’s aggressive push to embed AI deeply into core military operations and decision-making systems.

It makes Google the third major AI company to reach such a deal in recent weeks, following OpenAI and Elon Musk’s xAI.

The arrangement allows Google’s AI to be deployed in all lawful scenarios. It was structured as an amendment to an existing contract and includes language stating the tools are not intended for domestic mass surveillance or autonomous weapons.

OpenAI secured similar provisions, although AI and legal experts question whether the clauses are legally binding.

The move stands in direct contrast to Anthropic, which refused the Pentagon’s preferred terms and demanded explicit guardrails against those uses, triggering a monthslong feud with the Trump administration.

More than 600 Google employees signed a letter Monday urging CEO Sundar Pichai to reject the partnership. Google had previously withdrawn from its 2018 Project Maven military AI initiative after similar employee backlash.

This latest arrangement marks a major escalation in Big Tech’s direct support for highly classified U.S. defense programs.