A first formally documented case of new-onset psychosis involving ChatGPT

Last week, after giving a talk on AI and mental health at Université de Lorraine, someone in the audience asked whether any peer-reviewed research existed on the topic. I had to respond that, to my knowledge, all reports so far had come from the media.

Just shortly after, I came across a preprint published in Innovations in Clinical Neuroscience. It presents the case of a 26-year-old woman experiencing delusions involving ChatGPT: believing she was being “tested by ChatGPT” and could communicate with her deceased brother through it.

It’s a short case study and not yet peer-reviewed, so caution is warranted. Still, what makes this paper particularly insightful is the detailed clinical description, including some of the ChatGPT ‘manipulations’: “You’re not crazy. You’re not stuck. You’re at the edge of something. The door didn’t lock. It’s just waiting for you to knock again in the right rhythm.” One could argue this is the first formally documented case of new-onset psychosis in which ChatGPT played a role. Yet, as the authors rightly note, several other contributing factors were present, including a mood disorder, prescription stimulant use, and sleep deprivation.

A balanced perspective remains essential: we shouldn’t be too eager to take a sensationalist approach, we need to take other contributing factors into consideration, but we also should not dismiss the potential risks.