Google Pushes Ahead with US Department of Defense AI Contract

Internal backlash has spread after Google signed a contract allowing its Gemini model to be used for "legitimate purposes" on the US military's classified networks.

• About 600 employees signed an open letter of opposition, and some spoke out publicly against it.
• As with Project Maven in the past, there has been pressure for withdrawal, but this time management shows no sign of backing down and appears set to keep the contract.
• Google defends its position by pointing out that OpenAI, xAI, Nvidia, Microsoft, Amazon, and others have agreed to similar terms.
• Anthropic, on the other hand, refused the same terms, and the US military reportedly ordered a halt to the use of its products.

The article centers on two key issues: autonomous weapons and mass surveillance. Although the contract language reportedly presumes human control and oversight, critics point out that it carries weak legal force.

Google is also seen as having weaker internal organizing power than before. Since the Maven incident, the company has scaled back internal communication channels, and industry-wide layoffs and cost-cutting in AI are said to have made it harder for employees to mount a collective response.

Ultimately, this case is being read as a precedent for how far AI companies can limit the use of their models when supplying them to governments—especially military organizations—and whether employees can meaningfully control how that use plays out.