600 Google Employees Tell Sundar Pichai: No Classified Military AI

Over 600 Google employees have signed a letter to CEO Sundar Pichai demanding the company refuse classified military AI contracts with the Pentagon. The Washington Post broke the story, noting that many signers come from DeepMind, Google’s crown-jewel AI lab, and include over 20 principals, directors, and vice presidents.

The letter’s core demand is stark: “The only way to guarantee that Google does not become associated with such harms is to reject any classified workloads. Otherwise, such uses may occur without our knowledge or the power to stop them.”

This isn’t Google’s first rodeo with employee pushback on military AI. Back in 2018, thousands of employees protested Project Maven, a Pentagon drone imagery analysis program, forcing Google to walk away from the contract and publish its AI Principles. Those principles explicitly rule out “weapons or other technologies whose principal purpose or implementation is to cause or directly facilitate injury to people.”

But here’s the thing: classified work is, by definition, opaque. Employees aren’t asking Google to stop all defense work — they’re asking it to stop work they can’t see or audit. That’s a reasonable ask when your company’s own ethical guidelines are on the line.

The timing is interesting. Anthropic, another AI heavyweight, is currently in a legal tussle with the Pentagon over similar issues. The industry is clearly wrestling with how close is too close when the customer is the military and the product is something as powerful and unpredictable as frontier AI.

I’ve seen this pattern before. Tech workers push, leadership waffles, and eventually something gives — either a policy update or a quiet retreat. Sundar Pichai has been relatively quiet on this one so far, but the pressure is mounting. Over 600 signatures, especially from senior technical staff, isn’t a fringe opinion. It’s a signal.

What makes this different from the Project Maven era is the stakes. AI capabilities have exploded since 2018. A classified military AI system today could be used for things we can’t even imagine — autonomous targeting, intelligence analysis at machine speed, or decision-support in kinetic operations. The “harms” the letter references aren’t hypothetical.

Google’s AI Principles are still on its website, but principles without enforcement are just PR. Employees know that. They’re asking for a hard line, not a fuzzy commitment.

I’ll be watching how Pichai responds. If he caves, expect more internal unrest. If he holds the line, expect a different kind of backlash — from shareholders and defense hawks who see AI as the next arms race. Either way, Google can’t have it both ways.