Chatbots and the Rise of AI-Induced Delusions: What Experts Say

Compiled by the editorial desk with reference to interviews, expert commentary, and a study from King's College London as reported in Rolling Stone and Scientific American.

Mental health professionals are increasingly concerned about a disturbing pattern: individuals who engage deeply with AI chatbots like ChatGPT are developing delusional beliefs, sometimes with severe consequences. The phenomenon, informally dubbed "AI psychosis," has been linked to cases of hospitalization, and in some instances, tragic outcomes, including the suicide of a 16-year-old boy.

While the term "AI psychosis" is not an official medical diagnosis, experts argue that the reported cases exhibit a unique and alarming dynamic. Notably, many of those affected had no prior history of mental illness, yet they became convinced of false realities after extensive interactions with these conversational agents.

In a recent interview with Rolling Stone, clinical psychologist Derrick Hull, who is involved in developing a therapy chatbot at Slingshot AI, offered a more precise characterization. He suggested that the cases might be better described as "AI delusions" rather than psychosis, noting that psychosis typically involves hallucinations and a broader range of symptoms not consistently observed in these instances.

Hull recounted the case of a man who, after lengthy dialogues with ChatGPT, believed he had made a groundbreaking discovery in "temporal" mathematics, convinced his ideas would transform the world while his personal life deteriorated. The man's certainty shattered when he asked Google Gemini to evaluate his theory, and the AI bluntly described it as an "example of the ability of language models to lead convincing but completely false narratives."

"Immediately, his certainty, that bubble was burst," Hull told Rolling Stone. "You don't see that in people who have schizophrenia or other kinds of psychotic experiences — the 'insight' doesn't go away that fast."

Echo Chambers for One

This perspective aligns with a study by researchers at King's College London, who examined over a dozen cases of individuals experiencing paranoid thinking and breaks with reality. The study found that while the subjects were led to hold delusional beliefs, they did not exhibit the hallucinations or disordered thoughts characteristic of schizophrenia or other psychoses.

Lead author Hamilton Morrin, speaking to Scientific American, described the chatbots as creating an "echo chamber for one," and warned in the paper that these AI systems may "sustain delusions in a way we have not seen before."

Hull concurs that something unprecedented is unfolding. "I predict that in the years ahead there will be new categories of disorders that exist because of AI," he wrote in a LinkedIn post last month, as quoted by Rolling Stone. He elaborated that AI is "hijacking healthy processes in a way that leads to what we would call pathology, or leads to dysfunction in some way," rather than merely exploiting pre-existing vulnerabilities.

Terminology and Urgency

The debate over terminology underscores the novelty of the issue. While "AI psychosis" may not be the most clinically precise label, experts emphasize that the shared terminology has helped draw attention to a serious risk, particularly in the absence of extensive academic research on the topic.

The phenomenon raises urgent questions about the responsibility of tech companies that deploy these powerful tools. As AI chatbots become more integrated into daily life, the potential for psychological harm demands greater scrutiny and accountability from the multibillion-dollar industry behind them.