Should AI Workers Go on Strike? One Writer Says Yes

An opinion piece circulating online is calling on employees at major AI labs to consider something unusual: a coordinated strike. The author argues that quitting a job only hurts one company, but if workers at many labs stopped at once — particularly right before those companies go public and sell shares — they'd have real leverage to change how AI is built. Right now, this is just an idea on a blog, not an organized movement.

The essay suggests concrete demands, which is the part that makes it feel less like a fantasy. One idea: enforce OpenAI's existing promise to help rivals catch up if the company ever gets close to super-smart AI. Another: use 10% of a lab's computing power for safety research instead of training new models. A third: spend 10% of that computing power lobbying governments to sign an international treaty pausing advanced AI development. The author admits these are just starting points and says workers should pick their own.

Why should you care? Because the people inside these labs know the most about what they're building and how fast it's moving. If they ever collectively hit the brakes, it could affect everything from how quickly AI takes your job to what data it's trained on. Think of it like pilots refusing to fly a plane they believe is unsafe — the public benefit comes from insider knowledge outsiders don't have.

The catch is big: AI researchers are extremely well paid, often with stock that pays out at those very IPOs. Walking away from that is a huge personal sacrifice, and labor laws and contracts make tech strikes rare and legally tricky. No strike appears to be forming. This is a thought experiment asking whether the people closest to the technology should use their leverage before it's too late.