Google has given the Pentagon permission to use its AI models in classified military operations, thus joining OpenAI and xAI that They had already signed similar contracts previously. The AI majors are joining the US military apparatus. Anthropic is left alone.
what has happened. Google signed an agreement with the defense department last year worth 200 million dollars. This contract allowed the use of Google Cloud infrastructure and AI tools. The news now is that Google has given permission for the Pentagon to use the firm’s AI models in classified systems, for “any legitimate government purpose,” according to the New York Times.
Why it is important. Google’s decision to allow the Pentagon to use its technology contrasts with the case of Anthropic, that ended up being blacklisted for refusing to eliminate safeguards against autonomous weapons and domestic surveillance. This set a precedent: accept the conditions or be excluded from the government market.
Opacity. Speaking to the New York Times, a Google spokeswoman said the company remains committed “that AI should not be used for domestic mass surveillance or autonomous weaponry without appropriate human oversight.” However, they have not given specific details of the agreement and the fact that it will be used in “classified work” makes us wonder if Google really maintains some control or if it is simply an empty statement.
Employees against. At least 560 Google employees have signed an open letter to its CEO, Sundar Pichai, in which they ask him to reject the agreement. Employees argue that AI should be in the service of humanity, not for military purposes that include lethal autonomous weapons and mass surveillance. “The only way to ensure that Google is not associated with such damages is to reject any classified workloads. Otherwise, such uses could occur without our knowledge or ability to prevent them.”
New principles. Internal resistance at Google to the use of AI in military contexts is something that has been going on for a long time. In 2018, several employees resigned in protest to Google’s participation in Project Maven, the military program that used AI to identify people and objects to improve drone attacks. Google ended up leaving the project and pledged not to work on AI for weapons, but in 2025 that clause disappeared from its AI principles. Things have changed radically: today, big tech is more aligned with the US military than ever before.
Image | Xataka
In Xataka | Anthropic faced a long winter on the US “blacklist”: justice has saved it on the horn
GIPHY App Key not set. Please check settings