Psychologists warn about AI chatbots in mental-health support

The British Psychological Society poll included 197 psychologists, counsellors and psychotherapists.[1] Practitioners said they had seen a 49% increase in patients using AI for support over the previous 12 months.[1] Seven in 10 participants were somewhat or very concerned about clients using AI for mental-health support.[1] About three-quarters said AI was negatively affecting clients’ wellbeing.[1] These findings describe the views and observations of those surveyed; the evidence provided does not establish how common AI use is among all people seeking mental-health support.[1]

Nearly nine in 10 survey participants raised concerns about the lack of clinical oversight of AI chatbots, and 87% worried that AI would provide inaccurate information.[1] Eighty percent were concerned that clients were becoming overly reliant on AI.[1] One practitioner said they were concerned about not knowing how AI worked or generated its answers.[1] Another suggested that people might turn to AI because it feels safer than building trust with another person, potentially increasing isolation and reducing social contact and skills.[1] These are practitioners’ concerns, rather than evidence that every user experiences these effects.[1]

BPS president Dr Laura Dean said practitioners on the front line had cause for concern about people using AI for mental-health therapy, and that the risk of open-access, unregulated chatbots giving harmful advice remained real.[1] She said safeguards were needed to protect professionals and the people they support.[1] Dean also said AI could not replace the human element in mental-health care, despite its benefits, and called for more research and input from practitioners and AI users on safe integration with human-led therapy.[1] Separately, a survey of 2,000 children aged 11 to 17 found 43% had used AI for advice or emotional support.[1]