'AI Psychosis' in Corporations: Box CEO on a Wave of Reckless Swaps for AI Agents

Executives who decide that AI can replace your job often have no idea what you actually do. That is the paradox Box founder Aaron Levie pointed out, and he called it 'AI psychosis'.

When confidence outruns knowledge

Levie sees a worrying trend among his peers: managers are pushing to integrate AI into production without working out which tasks their teams actually perform. Instead of digging into the details, which takes time and effort, executives treat AI agents as a cure-all. The result is predictable: wrong replacements, lost context, derailed projects.

This is not a philosophical musing. It is already happening in real corporations, where employees learn about decisions to replace them not from their bosses but from the news online.

ClickUp as a bellwether

ClickUp recently laid off 22% of its team, citing AI agents as the reason. That shows how serious the wave is. But it is not an isolated case: according to TechCrunch, tech-sector layoffs in 2026 have already nearly reached the volume of all of 2025, even though the year has only just begun.

The statistics are both impressive and alarming. An industry that promised to create jobs and new opportunities is moving in the opposite direction. Companies are cutting staff in successive waves under the pretext of automation.

• The wave is happening in a rush, with no integration strategy
• Companies are chasing cosmetic cost reductions
• Management does not dig deep enough into the details of job roles
• Decisions are often driven by investment cycles rather than real need

Behind the scenes of these decisions

Venture investors demand a path to profitability as soon as possible. On paper, AI looks like a magic wand that will solve every problem. Company leaders look at neighbors who have already started deploying agents and fear falling behind. Instead of asking 'which tasks can we safely automate?', they ask 'how much will it cost us to give up on people?'.

This is psychosis in the classic sense: losing touch with reality in service of an idea. Managers watch the share price chart, not how their team works. They know AI can 'write code' and 'answer questions', but they don't understand why a human is needed for code review, team building and planning.

What this means in practice

If you are a developer, designer or analyst, you should know that the wave of replacements is often built on incomplete information about what you actually do. The best companies will most likely be those that do not simply adopt AI but understand its limits and risks.

For the companies themselves, 'AI psychosis' can end in long-term damage: loss of top talent, lower product quality, organizational chaos. History shows that the cost of hasty restructuring is often higher than the modest payroll savings promised to investors.