The tech industry is in the grip of a peculiar affliction: AI psychosis. That's the term coined by Box CEO Aaron Levie to describe a growing disconnect between top executives and the actual capabilities of artificial intelligence. In a viral post on X, Levie argued that CEOs are uniquely prone to this condition because they are too far removed from the nitty-gritty work that still requires human judgment. They see a prototype, generate a contract using an AI tool, and immediately jump to the conclusion that entire workflows can be automated by agents. But they never have to review code, catch bugs, or train models on idiosyncratic business rules.
Levie is not anti-AI; he's an active angel investor in AI startups and frequently posts about the future of headless software. Yet his diagnosis rings true. The first five months of 2026 have seen nearly as many tech layoffs as all of 2025: 115,430 people from 152 companies, compared to 124,636 from 275 companies last year, according to Layoffs.fyi. Many of these cuts are justified by CEOs as AI-driven efficiencies. But critics call this AI washing – using the hype to disguise other business decisions.
One stark example is ClickUp CEO Zeb Evans, who proudly laid off 22% of his workforce after deploying about 3,000 AI agents internally. He claims the move wasn't about cost reduction but about building a '100x org' where humans merely supervise AI output. However, the productivity data doesn't support such grand visions. A meta-analysis in UC Berkeley's California Management Review found no robust link between AI adoption and aggregate productivity gains. Research from the National Bureau of Economic Research noted a 'productivity paradox' where perceived gains exceed measured ones. And MIT researchers concluded that agents still fall short of human quality in many tasks, projecting that by 2029 they might reach 80%–95% success rates on text-related tasks – still not outperforming humans until a few years later.
The Harvard Business Review pointed out another bottleneck: when everyone uses AI to produce more output, executives become the chokepoint, needing to authorize a flood of work. Without proper grounding, CEO AI psychosis will likely lead to organizational chaos, not nirvana. Levie's advice is simple: CEOs must use AI 'a ton' to truly understand its limits and potential, and come out with an appreciation for both the upside and the real work. That wisdom seems to be lost on many leaders today.
Source: TechCrunch News