An NHS boss has joined mental health professionals in warning people with mental health problems against using artificial intelligence (AI) for support over the Christmas period. Dr Adrian James, national medical director for mental health at NHS England, said the technology cannot be “relied upon for everything and in some cases can be dangerous”.
He emphasised that it is “vital” that people know they can turn to the health service over the festive period, either by calling 111, referring themselves to talking therapies online, or by seeing their GP. The warning follows a survey published last month, which showed that more than a third of adults had turned to AI chatbots for help with their well-being.
Dr James said: “As a psychiatrist, I’ve seen an increase in the number of vulnerable patients turning to AI chatbots for mental health support over the last year.” In November, a poll of 2,000 people by the charity Mental Health UK found 37% have used an AI chatbot to support their mental health or wellbeing. When asked why they had turned to AI, around four in 10 people said it was down to ease of access, while almost a quarter cited long waits for help on the NHS.
While the NHS boasts that 95% of patients presenting with mental health issues should have a first appointment within 18 weeks of referral, wait times after an initial appointment can go for much longer. Dr James added: “During the festive period, I know Christmas can affect mental health in lots of different ways, whether it be financial pressures or feeling isolated, so it is vital that people know that they can turn to the NHS for help.”
He warned that the vast majority of AI chatbots do not have access to a user’s mental health history, cannot fully understand the nuances during a serious mental health situation, and can give completely wrong advice. “But my biggest worry is for those users who are at risk of losing touch with reality,” he said. “During an episode of psychosis, people are at higher risk of self-harm and suicide, and chatbots have an in-built preference to agree while lacking the sophistication to pick up on and to challenge problematic thoughts – this could lead to potentially dangerous situations.”
AI-induced psychosis is a new phenomenon being witnessed across the world. A series of suicides in America have been linked to AI chatbots, where several people took their lives after extensive and obsessive discussions with the chatbots. Roman Raczka, president of the British Psychological Society, recently stated that “AI offers many benefits to society, but it should not replace the human support essential to mental healthcare.”
