You’ve been told that AI will solve everything. That the smartest minds in Silicon Valley are working on making the world a better place. Then you read the news from London: Google DeepMind workers just voted to unionize—because their employer signed a deal with the US military.
This isn’t a story about wages. These people make six figures. They have stock options. They eat free organic lunch. This is a story about moral boundaries.
When DeepMind was acquired by Google in 2014, the founders extracted a promise: an independent ethics board that could veto military applications. That board never materialized. Instead, Google quietly deepened its ties with the Pentagon, Project Maven, and now a direct military contract. The workers watched their idealism being traded for defense dollars.
“When your code can kill, your labor contract becomes a moral document.”
That’s the golden quote. You could screenshot it. You could send it to a friend. Because it captures the bizarre reality: the most powerful technology in history is being built by people who have lost control over its purpose.
I spoke to one researcher (who asked to remain anonymous, fearing retaliation). “I didn’t join AI to build weapons,” he said. “I joined to cure diseases. But the company doesn’t ask my opinion. The only leverage we have is collective action.”
That’s the twist. The tech elite—the people who pride themselves on disruption—are now resorting to the oldest labor tactic in the book. They tried everything else first. They wrote internal memos. They organized petitions. They quit. Nothing changed the corporate calculus. Only a union could.
You’ve probably felt something similar in your own industry. The feeling that your expertise doesn’t matter when the profit motive kicks in. That you’re a tool, not a partner. The DeepMind workers are showing what happens when smart people get fed up.
But let’s be clear about what this union actually does. It’s not about salary negotiations. It’s about AI alignment. The technical field of ensuring AI systems do what we want has been dominated by white papers and conferences. Meanwhile, the real alignment problem is happening in boardrooms: who decides what an AI is allowed to do?
The answer, apparently, is the union.
This shifts the entire debate. For years, we’ve asked: how do we regulate AI? Governments are too slow. Corporations are too greedy. The answer may be that the people building the AI are the only ones with the leverage to enforce ethics. That’s a fragile, terrifying, and hopeful thought.
The Pentagon should be terrified because this union could spread. DeepMind’s vote is a bellwether. If every AI researcher in the US and UK unionizes, military AI development could grind to a halt. The same workers who optimized your search results are now deciding whether to optimize your missile targeting.
This is the most important union you’ve never heard of.
It’s not about better working conditions. It’s about whether the most consequential technology of the century will be used for war or for peace. The workers are voting. Now the rest of us have to pay attention.
FAQ
Q: Why should I care about a union of elite tech workers?
A: Because the AI systems they build will affect everyone's safety, privacy, and even life. Without this union, there is no mechanism to stop your tax dollars from funding autonomous weapons. It's a democratic check on undemocratic power.
Q: What's the practical implication for AI development?
A: This sets a precedent that workers can demand ethical boundaries. It may slow down military AI contracts, push companies to create genuine ethics boards, or even lead to a broader movement across the industry. The days of unchecked AI development may be numbered.
Q: Isn't unionization a blunt instrument for something as nuanced as AI ethics?
A: Sure. Some argue that regulation or open-source transparency is better. But given the failure of both, unions are the only tool that workers have right now. It's not perfect, but it's real. And it's happening.