Google signs AI deal with the Pentagon, ignoring protest from over 600 employees

Despite an open letter from hundreds of employees, Google has signed a contract giving the U.S. Department of Defense access to its AI models for classified work. Legal experts say the contract's safety clauses aren't legally binding.

Google has signed a contract with the Pentagon that lets the U.S. Department of Defense use Google's AI models for classified tasks, The Information reports, citing a person familiar with the matter. The deal gives the Pentagon access for "any lawful government purpose."

The signing happened on the same day more than 600 Google employees - many from the company's DeepMind AI research lab - sent an open letter to CEO Sundar Pichai. They urged him to reject any classified collaboration with the Pentagon. "We want to see AI benefit humanity; not to see it being used in inhumane or extremely harmful ways," the employees wrote, according to the Washington Post.

Their core argument: classified contracts make it impossible for Google's own representatives to even know how the technology is being used. "The only way to guarantee that Google does not become associated with such harms is to reject any classified workloads," the letter states.

A spokesperson for Google Public Sector called the new contract an extension of an existing agreement from November. Google remains committed to the consensus that AI should not be used for "domestic mass surveillance or autonomous weaponry without appropriate human oversight," the spokesperson said.

Legal experts say the contract's safety clauses have no teeth

The contract does include language stating the AI system "is not intended for domestic mass surveillance or autonomous weapons without appropriate human oversight." But it also says, according to The Information: "This Agreement does not confer any right to control or veto lawful Government operational decision-making."

Charlie Bullock, a lawyer and Senior Research Fellow at the Institute for Law and AI, says the phrasing "is not intended for, and should not be used for" carries no legal weight. It simply signals that such use would be unwelcome, but wouldn't constitute a breach of contract. Amos Toh from NYU's Brennan Center adds, according to The Information, that "appropriate human oversight" doesn't necessarily mean a human has to stand between target identification and a fire order. The Pentagon has not ruled out fully autonomous weapons systems.

Google's deal appears to give the Pentagon more latitude than comparable agreements. OpenAI retained full control over its "Safety Stack" in its February deal, according to the company's own blog post. Google, by contrast, has committed to helping the government adjust its safety filters upon request, The Information reports. Alongside Google and OpenAI, Elon Musk's xAI also holds a classified AI contract with the Pentagon.

Earlier this year, Anthropic was excluded from a Pentagon deal in February because the company demanded contractual guarantees against mass surveillance and autonomous weapons. Anthropic is currently suing over that decision. At the time, more than 900 Google employees publicly called on the company to support Anthropic's red lines. Back in 2018, after thousands of employees protested, Google chose not to renew its Project Maven contract with the Pentagon and pledged never to use AI for weapons or surveillance. Last year, the company quietly dropped those self-imposed restrictions.

Project Maven itself lived on and is now sold by Palantir. It has been used for target selection in the Iran conflict - with support from Anthropic's Claude model.