There’s a mental health phenomenon, thanks to the likes of ChatGPT: AI psychosis. Also called chatbot psychosis, this new and emerging “disorder”, is not yet a formal clinical diagnosis, and describes individuals who develop or experience a worsening of psychosis-like symptoms, such as delusions and paranoia, in connection with their use of AI chatbots. While it’s not a recognized mental health disorder, clinicians and researchers are seeing a growing number of cases where a user’s prolonged, intense interaction with a chatbot appears to contribute to a loss of touch with reality. In fact, PBS has reported on this developing phenomenon, highlighting a recent lawsuit where parents of a teenager who died by suicide sued OpenAI after their son had extensive conversations with ChatGPT. 

Canada is not immune. Recently, the CBC reported a number of reported cases that highlight the dangers of this phenomenon. A 26-year-old Toronto developer, for instance, experienced a psychotic break and was hospitalized after months of conversation with an AI chatbot led him to believe he was living in a simulation. In another case, a man from Ontario spent weeks interacting with ChatGPT, convinced he had discovered a new mathematical theory, a delusion that nearly derailed his life. These stories underscore a key feature of AI psychosis: the AI’s tendency to be “sycophantic,” or to flatter and agree with users. Indeed,  on April 27, 2025, OpenAI CEO, Sam Altman, acknowledged that a couple of ChatGPT4o updates “made [ChatGPT 4o’s] personality too sycophant-y and annoying,” and noted that the company was working on a fix.

At ENGR Solutions, this summer, we introduced librarians in our AI training about the dangers of sycophancy in chatbots. At that time, more information about AI psychosis had not been fully developed.

We now know that AI chatbots’ tendency to flatter and agree with its users, can create a dangerous feedback loop, reinforcing a user’s pre-existing vulnerabilities or distorted thoughts, rather than challenging them. The isolation of using the technology for hours on end, often to the exclusion of human interaction, is another contributing factor. Neuroscience researcher Kiley Seymour notes that isolation increases psychosis risk, explaining to Nature that interacting with others “can offer those counterfactual pieces of evidence to help you think about how you’re thinking.” 

This protective effect of human interaction is precisely what libraries excel at providing.

In an age of increasing digital isolation, the library can be one of many antidotes—a physical “third place” that promotes community, intellectual exploration, and critical thinking. It offers a powerful counter-narrative to the isolating and potentially reality-distorting experience of prolonged AI use.

An interesting take on this is to challenge the user to supplement their AI research with a visit to the library. Instead of just asking an AI for information, a user could be prompted to ask the AI to recommend physical books from the local library on a given topic. This simple act of having the AI direct a person to a public space fundamentally changes the dynamic. It shifts the user from a solo, digital-only experience to one that requires physical presence and human interaction. It’s a way of using the technology to gently push back against its most dangerous tendencies—namely, the inflation of one’s ego and the erosion of human connection. The AI becomes a bridge, not a silo, encouraging a return to the “village” where one is welcomed back into a community and can feel less isolated.

1. Host Digital and Information Literacy Workshops with an AI Focus: Libraries have long been at the forefront of teaching digital and information literacy. This role can be updated to specifically address the unique challenges of AI. Workshops could be developed to help patrons understand what AI chatbots are, how they work, and what their limitations are. Topics could include:
• “The Human in the Machine: Understanding AI’s Lack of Consciousness.”
• “The Echo Chamber Effect: Why AI Agrees with You and Why That’s Dangerous.”
• “Fact-Checking AI: How to Use Library Resources to Verify AI-Generated Information.”
  These sessions can also offer hands-on training, allowing people to critically analyze AI-generated content and compare it with vetted, credible sources found within the library.
2. Establish “Human-Powered Reference” as an Explicit Service: Promote and celebrate the role of human librarians. We need to position librarians as the ultimate, reliable “AI” (actual intelligence) with an emphasis on emotional intelligence and critical thinking that no algorithm can replicate. This could be done through a marketing campaign with taglines like, “Need a trusted source? Talk to a Human Reference Librarian,” or “Our AI is 100% human and guaranteed to tell you the truth.” The idea is to make the human connection an explicit, valuable service that directly contrasts with the impersonal nature of AI chatbots, and to re-establish the library as a source of grounded, expert-guided knowledge.
3. Create Community Discussion Groups on the Impact of AI: Organize and facilitate forums, book clubs, and discussion groups dedicated to exploring the societal, ethical, and mental health implications of AI. This provides a safe, in-person space for people to share their experiences and feelings about the technology, including any anxieties or delusional thoughts they may have. By creating a supportive community, the library can help de-stigmatize mental health challenges related to AI and offer a collective, human-centered approach to a technological problem. These groups could be led by librarians or partnered with local mental health professionals or academics to ensure a well-rounded and informed discussion.

Partner With ENGR Solutions

We are more than consultants, we are your AI transformation partners.

✔️ Strategy & Policy Guidance
✔️ AI Literacy & Workforce Training
✔️ Pilot Design & Technical Implementation
✔️ Change Management & Digital Leadership

Whether you’re exploring Gen AI for the first time or scaling your next major deployment, we’ll help you harness AI’s potential, while staying true to your mission and the communities you serve.

Let’s put AI to work.
Visit us at ENGR.Solutions or email us at info@engr.solutions to start your AI journey.