Google employees fail to block AI partnership for classified Pentagon operations

Google leadership has moved forward with plans to develop AI systems for classified Pentagon operations, despite a letter from over 600 employees urging CEO Sundar Pichai to block the deal.

Unlike the 2018 employee protest against Project Maven—which successfully pressured the company to let its drone-video contract expire—this latest internal push failed to shift executive strategy.

The Pentagon recently confirmed that Google is among seven major firms, including OpenAI and SpaceX, now approved to handle high-level classified AI projects.

The tension has led to significant changes in Google’s internal culture. Monthly "all-hands" meetings, once known for their transparent and direct nature, are now moderated through an AI tool internally known as "Ask."

This system allows moderators to summarise and reword employee submissions before they are presented to executives.

Staff members report that pointed questions regarding sensitive topics, such as Google’s involvement with ICE and Customs and Border Protection, are being reframed into more general inquiries about "government agencies."

The controversy is further fueled by Project Nimbus, a $1.2 billion cloud agreement signed with the Israeli government in 2021.

Recent reports and whistleblower declarations suggest that Google has assisted the Israel Defence Forces in developing AI-driven object-identification capabilities.

Software engineers have expressed frustration, claiming they only learned of these military applications through news sources.

As major tech employees from Google, Amazon, and Microsoft begin to organise across company lines, many describe a "real and impending sense of doom" regarding the ethical trajectory of the tools they are building.