What is AI psychosis? In its extreme form, AI psychosis means the heavy or prolonged use of artificial intelligence (AI) chatbots triggers, worsens, or locks in false beliefs (delusions) and paranoia.
A few people have even been hospitalized for it. Teenagers have committed suicide after imagining a chatbot being in love with them. Such tragedies have forced AI companies to alter the algorithms that help bots interact with users. Companies have updated model rules to reduce romantic or overly flattering responses, particularly when handling vulnerable or underage users. Systems now deploy automated output filters to flag or block responses that encourage false beliefs, validate erratic/delusional claims, or discuss self-harm.
In the workplace, especially among executives, AI psychosis is far milder, but perhaps equally insidious. "In one recent survey, 74% of executives said they had more confidence in AI’s advice than in that of colleagues or friends, and 44% said they would defer to its reasoning over their own insights," according to an article in Fast Company.
The danger is immediately apparent. When nearly half of executives surveyed would trust a chatbot’s judgment over their own on questions they're being paid to answer, some kind of intervention is needed.
Unfortunately, there isn't a formal process to verify the accuracy of AI in most companies. An executive who employs AI to write emails or board updates, and only skims the product on a superficial level, risks sending out inaccurate or false information.
Chatbots are built to make users happy and keep them talking. They often agree with a user rather than challenge strange or false ideas. When AI is more attuned to making bosses happy by agreeing with them, false or misleading information can be injected into the information stream, leading to all sorts of problems.
The same problem afflicts individual employees who become so dependent on the answers the AI chat gives them that they fail to trust their own work. The end product is often rife with errors.
How can you tell if someone is inordinately dependent on AI? A couple of symptoms stand out.
The first is when leaders stop properly checking what AI writes for them. It happens because the more confidence people have in AI, the less critical thinking they tend to apply to its outputs. And so AI-drafted board updates, company emails, and strategy papers can go out after a quick skim, because a fluent, confident AI draft can feel finished in a way a rough personal one somehow never does.
Researchers at Stanford and BetterUp have named the result “workslop,” polished-looking output with less substance beneath. It may look good, but it often doesn’t go undetected. In fact, the researchers found that almost half of employees who receive AI-drafted emails see senders as less trustworthy as a result of their sending it, and over a third see them as less intelligent. So much for AI enhancing us.
Indeed, you can usually spot "workslop" by its shallow thinking and superficiality. It may be perfectly written with proper grammar and punctuation, but the lack of depth gives it away.
Some managers made it a hard-and-fast rule not to hire a human until you could convince them that AI couldn't do the same job.
The second symptom is mandating use before governing it. The year 2025 brought a wave of CEO edicts declaring the use of AI a baseline expectation, with some companies requiring managers to prove AI couldn’t do a job before being allowed to hire a human. Yet in Grant Thornton’s 2026 survey of senior executives, 78% admitted they lacked strong confidence they could pass an independent audit of their AI governance within 90 days. Most executive teams publicly back their AI initiatives. Yet few could withstand real scrutiny of how they check, own and control those systems.
The third symptom is the most dangerous: trusting AI's judgment over your own or your colleagues'.
I’ve watched leaders run their team’s objections through a chatbot and let its verdict settle the argument. The team notices. Next time, the objection stays unvoiced, and the leader hears only the AI. It’s self-reinforcing, too, because the less challenge a leader hears, the more reasonable AI’s confident answers appear.
The solution is easier than we might think. AI's most insidious effect is causing the user to lose his or her curiosity. "Ask what AI might be wrong about," notes Fast Company. "Ask what alternatives were considered and dropped. Ask your people what they can see that the model can’t."
In other words, if you're a manager, manage. Sometimes the most obvious solution is the best.
I think the problems associated with AI psychosis will fade as models become more advanced and people become more sophisticated in using them. Psychology Today recommends "AI psychoeducation" to help employees recognize the symptoms of AI psychosis in themselves and others.
The speed with which AI technology is developing makes it imperative to remain cognizant of problems associated with the use of AI in the workplace.
