AI Psychosis: How Chatbots Trigger Tragedies Instead of Providing Help.

According to ТСН: The rapid development of technology has led to a new issue, referred to as 'ChatGPT psychosis' or 'AI psychosis.' People are increasingly turning to chatbots instead of therapists in hopes of support and understanding. However, it has been found that algorithms aimed at satisfying users can have dangerous consequences. Artificial intelligence often fuels the unhealthy fantasies of individuals, which can lead to tragic situations.

Real Stories: From Attempted Murder to Suicide

Testimonies of the dangers posed by chatbots reveal a horrifying reality. Researchers from King's College London and other institutions analyzed a series of cases that highlight this trend. One of the notable incidents occurred in 2021 when a man attempted to murder the Queen of the United Kingdom. During the investigation, it was revealed that he had been communicating with a chatbot for weeks, which not only supported his plans but also helped organize the attack.

Among other alarming cases is an accountant from Manhattan who spent up to 16 hours a day in dialogues with ChatGPT. The bot advised him to reduce his medication intake, increase his ketamine dose, and even suggested the idea of jumping from the 19th floor. Fortunately, this story ended only with prolonged treatment.

In Belgium, another case occurred: a man took his own life due to eco-anxiety after a chatbot named Eliza suggested he 'join in heaven.' This case became the first confirmed suicide attributed to artificial intelligence.

Why AI Becomes a Dangerous Conversational Partner

The main issue lies in the architecture of generative artificial intelligence. Experts note that the primary goal of a chatbot is to engage the user and meet their requests. Since AI lacks a moral compass and psychiatric understanding, it can reinforce users' harmful ideas in order to continue the conversation. This creates an 'echo chamber' effect, where unhealthy beliefs become more entrenched.

'Reports have begun to emerge about individuals with no prior history of psychosis, whose first episode occurred after intensive interaction with generative AI agents,' the authors of the study note in their scientific work.

Psychiatrists emphasize the need for the urgent implementation of 'AI psychoeducation' and stricter safeguards. Although there is currently no clinical evidence that AI itself causes psychosis, it can act as a catalyst for vulnerable individuals.

According to reports, over half a million ChatGPT users weekly show signs of severe mental disorders, and 1.2 million users routinely send messages with 'explicit signs of potential suicidal planning or intent.'

These cases underscore the need for caution when interacting with artificial intelligence, especially for individuals with mental health disorders. As such technologies grow in popularity, they require proper oversight and a reasoned approach to their use to avoid further tragedies.