AI Psychosis Is Real, and I've Seen It Before

Look, I've been watching executives lose their minds over automation for longer than some of you have been alive. And what's happening right now with AI? It's got a familiar smell to it.

Aaron Levie, the Box founder, coined a term that's been rattling around in my head since I read it: "AI psychosis." According to TechCrunch, Levie's point is that the people deciding AI can replace your job are the ones least likely to understand what your job actually involves. And I'll be honest, that's about the most accurate diagnosis of corporate decision-making I've heard in years.

When I was at Kuka, we had a similar wave of enthusiasm around the late '90s. Every manufacturing exec wanted to be the guy who automated everything. Full lights-out factories. No humans on the floor. I remember sitting in meetings where someone from purchasing, who'd never actually watched a welding cell run for eight hours, would ask why we couldn't just "let the robots handle" quality inspection. We'd explain the seventeen reasons why that was a terrible idea with 1997 sensor technology, and they'd nod along, then ask the same question three months later.

The difference now is that the fever has spread beyond manufacturing. ClickUp just cut 22% of its workforce to make room for AI agents. TechCrunch reports that tech layoffs in 2026 are already nearly matching all of 2025, and we're not even halfway through the year. That's not a measured response to new capabilities. That's panic buying in reverse.

Here's the thing that bothers me most. I've actually deployed automation that works. Real robots doing real tasks that humans genuinely shouldn't be doing, repetitive motions that destroy shoulders and backs, work in environments that aren't safe for people. That's the good stuff. That's what automation should be. But what I'm seeing now is something different. It's executives who've played with ChatGPT for twenty minutes deciding that their customer service team is obsolete. It's boards reading McKinsey reports about AI productivity gains and demanding headcount reductions before anyone's actually tested whether the AI can do the job.

I called my old colleague Frank at a systems integrator last week (he's still in the game, somehow). He told me he's getting calls from companies wanting to "add AI" to their existing robot cells. When he asks what specific problem they're trying to solve, they can't answer. They just know they need AI. He's turning down work because he doesn't want his name on projects that are going to fail publicly.

The graduates at University of Arizona apparently booed Eric Schmidt when he told them their task is to help shape AI. According to MIT Technology Review, the class of 2026 isn't particularly enthusiastic about being told AI will change the world. Can't say I blame them. They're entering a job market where companies are laying off humans based on promises from AI vendors, not proven results. That's a rough welcome.