Google’s Confidential Agreement with Pentagon on AI
The use of artificial intelligence is not only of interest to businesses, but also to the military. While Anthropic and the Pentagon are in conflict over disagreements on the limits of the use of generative AI by the US military, OpenAI has formalized an agreement. Now, Google reportedly also has a confidential agreement with the Pentagon on AI. This information was revealed by Reuters, citing sources close to the matter. Essentially, the Pentagon now has the ability to utilize Google’s AI for all legally permissible purposes. Additionally, to meet the criteria of the US military, Google would need to review the security parameters of its AI to make it more permissive.
The contract reportedly specifies that Google’s AI should not be used “for domestic mass surveillance or autonomous weapons (including target selection) without appropriate human oversight and control.” Details shared by Reuters have not been confirmed. However, according to the agency, a Google representative reportedly told The Information that this new contract is an amendment to an existing agreement between the Pentagon and Google.
Internal Disagreement Source
It seems that a large number of Google employees do not agree with the military usage of the company’s AI by the Pentagon. An article in The Washington Post (via The Verge) reported that over 600 employees signed a letter addressed to CEO Sundar Pichai opposing the project.
Furthermore, the organizer of this petition indicated that many signatories are employees of DeepMind, Google’s AI lab, which develops the Gemini models. The list of signatories also includes directors and vice presidents.
Google is in a delicate position as Washington has already sanctioned Anthropic for not agreeing to the US government’s conditions regarding the military use of its AI.
Our Opinion
These agreements between the Pentagon and AI labs show that artificial intelligence is no longer just a productivity tool but also a defense tool. In France, the Ministry of Armed Forces has already established a framework agreement with Mistral AI. “Our solutions will be deployed on French infrastructure, ensuring complete control of data and critical technologies,” Mistral explained in January, according to a previous Reuters article.
- Similar to OpenAI, Google has a new agreement with the Pentagon allowing the US military to use its AI
- Google clarifies that it is an amendment to an existing agreement
- Rumors suggest that 600 Google employees have opposed this project
/2025/12/24/zidane-694bbd63de5c4791782488.jpg "2026 World Cup: Jaw fracture for Algerias goalkeeper Luca Zidane, a month and a half before the competition")
/image%2F1371318%2F20260427%2Fob_daa03c_bc5c6b79-11e0d4b61ba58497c9eeac38c017d.jpg "We watched – at the cinema")
