Chatbot psychosis

“Chatbot psychosis”, a term not yet found in the clinical literature but increasingly discussed, raises the question of whether AI chatbots can create a new mental health risk.

The concept, which has recently entered public debate, suggests that intense and prolonged interaction with AI chatbots may distort a person’s perception of reality. Users who are psychologically vulnerable may find it harder to tell the real from the virtual, and this may trigger mental health risks such as anxiety, paranoia and psychotic symptoms.

Uncontrolled and improper use creates new risks

AI solutions offer important opportunities in healthcare, from early diagnosis to personalized treatment. However, uncontrolled and improper use of these technologies also brings new risks. What is called chatbot psychosis can emerge when safe design, ethical boundaries and user awareness are missing. In healthcare, these solutions need to be implemented with human psychology and safety in mind.

Managing the risks requires joint action

Managing these risks requires broad collaboration, from technology developers to regulators:

• The clinical and psychological effects of chatbots should be monitored regularly through independent research.
• Developers should add safety layers to their systems and build mechanisms that detect risky usage patterns and direct users to appropriate support.
• Regulators should put ethical and safety standards specific to this area into effect.
• Technology companies should place transparency, auditability and user safety at the center of their work.

Although chatbot psychosis is not yet part of the clinical literature, it is a phenomenon that needs careful monitoring. So that such risks are not overlooked alongside the opportunities AI offers, technology companies, health professionals and regulators need to act together.

For principles on the safe use of general-purpose chatbots in healthcare, see the guide General-purpose chatbots and clinical decision support.