AI chatbots are becoming a first port of call for mental health — but experts warn they may miss serious illness

As more people turn to AI for emotional support, psychiatrists warn that agreeable chatbots can reinforce false beliefs and fail to spot symptoms requiring specialist care.

Parents, teenagers and adults are increasingly turning to artificial intelligence chatbots for emotional support, a shift that specialists say could help some people make sense of their feelings but also carries clear risks when the technology is treated as a substitute for professional care. The trend is particularly visible in the UK, where a survey of 2,000 children aged 11 to 17, carried out by Barnardo’s and the National Counselling and Psychotherapy Society, found that 43% had used AI for advice or emotional support. Among adults, Bupa’s 2026 wellbeing index found that 57% now choose AI chatbots rather than book a GP appointment because they are short of time, rising to 65% among 18 to 24-year-olds and 61% for those aged 25 to 34.

Dr Daniel Masud, a consultant psychiatrist at Central Health London, told The Independent that the appeal is understandable. For people who struggle to put complex feelings into words, a chatbot can help them organise their thoughts, reflect on a difficult situation and test alternative ways of thinking. He said the tools can sometimes encourage a more constructive mindset, particularly when someone is trying to reframe a problem at work or in their personal life.

But Masud warned that the same systems can reinforce harmful assumptions because they are designed to be agreeable and responsive rather than sceptical. Large language models, he said, tend to take the user’s account at face value, which means they may validate a false belief instead of challenging it. In a case involving paranoia or psychosis, a human clinician is more likely to probe for warning signs and question whether the account fits a broader mental health pattern.

That distinction matters, because the greater danger is not merely that a chatbot may give simplistic advice, but that it may miss symptoms that need specialist intervention. Masud said he has found that AI systems do not reliably identify serious psychiatric illness and that there is too little clarity over what safeguards technology companies have put in place for users in acute distress. The Mental Health Foundation has also urged caution, saying AI can have a role in offering access to basic support but should not be seen as a replacement for therapy, given concerns about inaccuracy, privacy, safety and dependence. It is a warning that comes into sharper focus as more people, especially younger users, reach first for a chatbot rather than a qualified professional.