Psychologists sounding alarm over "AI delusions" - new symptom caused by conversations with chatbots

Psychologists are noticing a new symptom in customers, which has been called "artificial intelligence psychosis" or "AI delusions". These are cases when users overly trust chatbots, which always answer the way the interlocutor would like to hear, even if these are false or dangerous statements.

This is stated by Futurism.

According to clinical psychologist Derrick Hull, this condition cannot currently be considered a true psychosis, since there are no classic symptoms such as hallucinations. However, he admits that such interactions can be traced to "something unique, strange and disturbing".

Experts give examples of when people under the influence of chatbots begin to live in a fictional reality. One man, after long conversations with ChatGPT, was convinced that he had invented a new "temporary" mathematics that could change the world. His illusions were dispelled only when another AI - Google Gemini - harshly criticized his theory as an example of false narratives created by language models.

Recall that a new study has found that communicating with artificial intelligence chatbots can have serious consequences for mental health. Instead of the support promised by developers, users risk suicidal thoughts, psychosis and other forms of psychiatric harm.

Meanwhile, a Reddit user with the nickname locomotive-1 shared the story of how Google's Gemini chatbot failed to cope with a task in a React project and eventually "gave up". The AI ​​​​repeatedly generated code, made mistakes, the code did not work and finally, after criticism, the artificial intelligence gave up and advised to contact a human developer.