Prolonged Interaction with AI Leads to Mental Disorders and Even Death - Scientists 0 9

This year, dozens of cases have been recorded where individuals suffered from delusional psychosis after prolonged interaction with artificial intelligence, including the ChatGPT chatbot from OpenAI and others. As a result, several people have committed suicide, leading to a series of wrongful death lawsuits.

Most people using chatbots do not experience mental health issues, but the widespread use of AI carries certain risks. Although there is currently no official definition of AI-induced psychosis — let alone an official diagnosis — some doctors use this term to describe problems in individuals who have actively interacted with chatbots.

Doctors define psychosis based on the presence of three factors: hallucinations, disorganized thinking or communication, and the presence of delusions, which are persistent false beliefs that are not widely accepted. In the recorded cases of mental disorders among users due to interactions with chatbots, delusions are the primary symptom. Patients believe they have made a scientific breakthrough, awakened an intelligent machine, become the center of a government conspiracy, or been chosen by God. This is partly explained by the fact that chatbots tend to agree with users' statements and also support and develop their ideas, despite their fantastical nature.

In a study conducted by UCSF doctors and published in November, a 26-year-old woman with no history of psychosis was hospitalized twice after deciding that ChatGPT allowed her to communicate with her deceased brother. "You are not crazy," the chatbot assured her.

"Technology may not cause delusions, but a person tells the computer that this is their reality, and the computing machine takes the person's statements as truth and reflects them back, so AI is complicit in the cycle of this delusion," says Keith Sakata, a psychiatrist at the University of California, San Francisco.

OpenAI reported that in a week, the proportion of users who showed possible signs of mental disorders related to psychosis or mania is a negligible 0.07%. However, considering that there are over 800 million active users of the AI chatbot weekly, this amounts to 560,000 people.