Box founder Aaron Levie has coined a startling term for the current state of corporate boardrooms: AI psychosis. He is calling out executives who are convinced AI can replace entire roles, despite having little understanding of what those jobs actually entail. This diagnosis highlights a growing disconnect between leadership and the reality of daily operations.
The timing of this critique could not be more relevant to the current tech landscape. ClickUp recently slashed 22% of its workforce, explicitly replacing people with AI agents. This move signals a shift from experimentation to aggressive cost-cutting based on automation assumptions.
We are only in May, yet tech layoffs in 2026 are already nearly matching the entire year of 2025. This acceleration suggests that the industry is moving faster than its ability to responsibly integrate these tools. Companies are betting big on AI before the technology has proven stable in production environments.
Levie's point cuts to the heart of a dangerous pattern in modern management. The further you are from the actual work, the easier it looks to automate. CEOs see tasks and tickets, not the judgment calls, context switching, and institutional knowledge that make someone effective.
This matters if you're using AI in your work right now. The gap between what AI can do and what leadership thinks it can do is creating real risk. Companies are making irreversible staffing decisions based on optimistic demos rather than production reality.
The term 'AI psychosis' might sound dramatic, but it captures something real. There is a disconnect between the people deploying AI at scale and the people who understand the work being automated. As the original outlet noted, that gap has consequences, and we are watching them play out in real time.
What this means for you: Do not assume your role is safe just because AI is not replacing it yet. Leadership may not understand the nuance of your work, which makes your value invisible to them. Protect your position by documenting the complex decision-making processes AI cannot replicate. Try this prompt with your AI assistant to audit your workflow: 'Identify three tasks in my weekly routine that require high-context judgment or institutional knowledge, and explain why an AI agent would struggle to perform them accurately without human oversight.' Use this analysis to demonstrate your unique value to stakeholders who only see task completion.
