Pentagon turns to Google AI for classified operations - Details emerge
Google risks conflict with its own developers over Pentagon cooperation (photo: Pexels)
Google has joined the list of Silicon Valley tech giants supplying AI models for the needs of the US military. A new secret agreement allows the Pentagon to use the company’s technologies for any lawful government purposes, according to Wall Street Journal.
According to sources, the contract includes the use of AI for mission planning and possibly target identification. The deal places Google alongside Sam Altman’s OpenAI and Elon Musk’s xAI, which are already cooperating with the US Department of Defense.
Interestingly, Google has agreed to adjust the safety settings and filters of its AI models at the government’s request.
Although the document states that the AI is not intended for autonomous weapons or mass surveillance without human oversight, Google does not have the right to veto military operational decisions.
Internal protests at the tech giant
Company employees have expressed serious concerns that their developments could be used for inhumane or extremely harmful purposes. In an open letter to CEO Sundar Pichai, they are demanding that the company отказаться from working on classified military tasks.
“The Pentagon is negotiating with Google and OpenAI to try to get them to agree to what Anthropic has refused. They're trying to divide each company with the fear that the other will give in. That strategy only works if none of us know where the others stand,” the letter says.
It is noteworthy that Google previously refused similar contracts (such as Project Maven in 2018) after mass employee protests. However, in 2025, Google’s parent company Alphabet lifted its internal ban on using AI for weapons and surveillance, citing national security needs.
Are top AI systems being militarized?
The Pentagon is actively involving leading AI labs in the development of classified digital tools. It is known that the US Department of Defense has signed agreements worth up to $200 million each with Anthropic, OpenAI, and Google.
Unlike Google, the startup Anthropic refused in early 2026 to weaken the safety barriers of its Claude AI. As a result, the Pentagon labeled the company as risky for the supply chain.
Google, on the other hand, chose the path of cooperation, considering providing API access to its commercial models a responsible approach to supporting national security.
