We have gotten used to thinking of AI psychosis as the result of an addiction to AI chatbots to the point of self-harm. The positive reinforcement built into chatbots leads to reinforcing psychosis, leading some users to suicide and others to criminal activity.

If you are having thoughts of suicide or self-harm, whether or not it’s chatbot-induced, we recommend contacting the Crisis and Suicide Lifeline at 9-8-8 or 988lifeline.org.

There is another kind of AI psychosis, however, that is not caused by using chatbots but by a delusional belief in their capabilities. This kind of AI psychosis was explained by Cory Doctorow, former editor of Boing Boing and a digital rights activist, in a recent issue of his newsletter, Pluralistic.

Doctorow discusses investor delusion — the belief that AI is going to generate profits even if it never has because it simply has to. He describes boss delusion — the belief that AI is going to enable you to fire all your employees or supercharge their productivity. You can make a case that AI psychosis is a form of corporate self-harm.

I’m going to make that case.

In making this case, I’m going to enlist the support of the “paper of record” for the addiction treatment industry, Better Health Business (BHB). The publication gives industry executives the opportunity to interview anonymously about what’s going on in the addiction treatment business.

The feature is called “BHB Confessions,” and the latest confession is a doozy. It comes from the former CEO of one of the largest nonprofit addiction treatment systems who recently retired after 50 years in the behavioral health field. BHB asked him, “What’s new,” and out popped this:

What I have seen that troubles me is all of a sudden, there’s this drive that you have to have AI. You have to have it if you don’t have it.

AI psychosis has come to the management of addiction treatment. “There are these things that come along that are driven by a lot of the consultants that really don’t play out,” the CEO tells BHB. For example, much of the information in healthcare is in silos that make it difficult, if not impossible, for AI to collect and analyze the data it needs to achieve breakthrough results.

Health insurers are using AI, but they do not want to share their data with governments or healthcare providers. So clinics end up dealing with patients without access to medical history, and governments are making policies without access to insurance company data to measure effectiveness.

The AI can come up with great ideas, but it can’t make the budget for them appear. Every penny of most budgets is already spoken for, and new ideas require new sources of funding with them. The funding, according to this CEO, comes from people who are passionate about projects, not from AI making the organization more efficient. 

“It would be most important to be able to build from the bottom up, not from the top down,” the retired CEO told BHB. You put the needs of the people using addiction treatment ahead of the desire of the organizations providing care to operate more efficiently. 

AI in addiction treatment to date is another top-down idea being imposed, whether it works or not, at the expense of what we know does work.

Written by Steve O’Keefe. First published August 31, 2026.

Sources:

“Confessions from a Retired Addiction Treatment CEO: The Industry Keeps ‘Recycling Solutions’,” Better Health Business, August 18, 2026.

“Three more AI psychoses,” Pluralistic, March 12, 2026.

Image Copyright: dragonimages.