Remember when Google employees revolted over Project Maven and the company promised not to use AI for weapons? That lasted about six years. Google is now providing the Pentagon with Gemini-powered AI agents for use by 3+ million military and civilian employees, marking a dramatic reversal in the company's stance on defense contracts.
The 2018 Project Maven protests were a defining moment in tech worker activism. Thousands of Google employees signed petitions, dozens resigned, and the company ultimately declined to renew the contract. Google then published AI Principles that specifically limited military applications. Those principles apparently had an expiration date.
According to Emil Michael, Under Secretary of Defense for Research and Engineering, the Gemini agents will initially operate on unclassified networks, with talks underway to expand to classified systems. The eight pre-built agents are designed to automate routine tasks: summarizing meeting notes, constructing budgets, reviewing proposed actions against defense strategy. Sounds harmless enough—until you remember that "routine tasks" in a military context include things like targeting decisions and operational planning.
The question is what changed. Did the technology evolve to the point where Google is comfortable with military applications? Did the geopolitical situation—rising tensions with China, wars in Ukraine and the Middle East—shift the company's calculus? Or did the internal resistance simply fade as employees who cared about Project Maven left and were replaced by a generation that sees Pentagon contracts as just another customer?
My guess: it's the money. The Department of Defense represents one of the largest technology buyers in the world, and competitors aren't sitting out. has major defense contracts. runs classified workloads. built its entire business on government work. If refuses to compete, they're leaving billions on the table while rivals strengthen their government relationships.
