The AI Psychosis: When Executives Automate Jobs They Don't Understand

The logic sounds clean and inevitable: artificial intelligence can automate tasks, so replace expensive human labor with cheaper AI agents. But according to Box founder and CEO Aaron Levie, this reasoning has a fundamental flaw. The executives making these decisions, he argues, are often the people least qualified to assess the work being replaced. Levie calls this phenomenon “AI psychosis.”

The Gap Between Perception and Reality

Levie’s critique cuts to the heart of a growing corporate trend. In recent months, companies like ClickUp have announced significant layoffs directly tied to AI adoption. The project management firm cut 22 percent of its workforce, explicitly citing the deployment of AI agents as the reason. Data from layoff trackers shows that tech layoffs in 2026 are already nearly matching the total for all of 2025, with AI displacement accelerating.

The problem is not automation itself. Automation has always eliminated some jobs while creating others. The danger lies in the disconnect between what executives believe AI can do and what the work actually requires. A manager who has never written code, designed a product, or managed a complex customer relationship may believe that an AI agent can replicate those functions. This gap between perception and reality creates a new class of business risk.

The ClickUp Case and Its Implications

ClickUp’s decision to replace 22 percent of its workforce with AI agents provides a concrete example. The company builds productivity software, a market where nuance and user context matter deeply. An AI agent can triage tickets or generate basic reports, but it cannot understand the subtle political dynamics of a client relationship or the creative leap required to solve an ambiguous design problem. By treating these roles as interchangeable with software, ClickUp may have saved salary costs while introducing operational blind spots.

Other companies are following a similar playbook. The scale of layoffs in 2026 suggests that many executives see AI as a lever for immediate cost reduction rather than a tool for strategic transformation. This approach mirrors the early days of outsourcing, when companies moved call centers offshore only to discover that customer satisfaction dropped and complexity costs rose. The difference this time is speed: AI deployment can happen in weeks, not months, and the consequences may compound faster.

What Practitioners and Decision Makers Should Watch

For professionals whose roles involve judgment, creativity, or relationship management, the message is clear: your value lies in the parts of your work that are hardest to describe in a prompt. Document those skills. Make them visible. For executives, the antidote to AI psychosis is humility. Before replacing a team, spend time understanding the actual workflow, the tacit knowledge involved, and the edge cases that an AI agent cannot handle.

The coming year will test whether companies that rushed to automate will suffer from degraded quality, slower innovation, or hidden costs. The most successful organizations will likely be those that treat AI as an assistant, not a replacement. They will augment human judgment rather than bypass it. The executives who understand this distinction will avoid the psychosis that Levie warns about. The ones who do not will learn the hard way that you cannot automate a job you do not understand.

Source: TechCrunch AI