Over 600 Google Employees Tell Sundar Pichai: No Classified Military AI

Another day, another internal revolt at a big tech company over military AI. This time it’s Google, and the stakes are higher than I’ve seen in a while.

Over 600 Google employees signed a letter to CEO Sundar Pichai, demanding the company block the Pentagon from using its AI models for classified purposes. The Washington Post broke the story, and the numbers are notable: organizers say many signers come from DeepMind, Google’s crown-jewel AI lab, and include more than 20 principals, directors, and vice presidents.

That’s not junior engineers grumbling on Slack. That’s senior technical leadership putting their names on the line.

The letter’s core argument is blunt: “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 is the same company that famously pulled out of Project Maven in 2019 after employee protests, then published a set of AI Principles that explicitly barred using AI for weapons or surveillance that violates international norms. Those principles are now the elephant in the room. If Google is entertaining classified military contracts, it’s hard to see how that squares with its own stated rules.

What’s especially interesting here is the timing. Anthropic, another major AI player, is currently locked in a legal battle with the Pentagon over similar issues. The AI industry is getting squeezed between lucrative government contracts and the ethical lines drawn by their own employees and founders.

Let’s be real: classified military work is a black box. Once the data and models go behind that wall, nobody outside the classified program knows what’s happening. The employees’ fear isn’t just abstract — it’s practical. If Google builds a general-purpose AI model that gets used for targeting, drone coordination, or intelligence analysis, and it’s classified, the engineers who built it may never find out until it’s too late.

Google hasn’t responded publicly yet, but Pichai now has a decision to make. Say no, and risk losing a massive Pentagon contract — the kind that keeps Wall Street happy. Say yes, and risk a full-blown talent exodus from DeepMind, where many researchers explicitly joined because of the AI Principles.

I’ve seen this movie before. Microsoft went through it with the JEDI contract. Amazon employees pushed back on Rekognition being used by police. But Google’s situation is trickier because their AI is genuinely state-of-the-art, and the military knows it. The Pentagon isn’t going to walk away just because some employees are uncomfortable.

What I find frustrating is how predictable this is. Every few years, a tech company gets caught cozying up to the military, employees freak out, the CEO makes a vague promise, and then a few years later, the same fight happens again. The business incentives haven’t changed. The military has money, prestige, and a sense of urgency that commercial clients don’t. Google’s leadership has to decide whether their AI Principles are real policy or just PR.

I’m watching this closely. If Google caves on this, it sets a precedent for every other AI lab facing similar pressure. And if they hold the line, it might actually mean something.