AI cannot give the empathy and trust at the heart of mental health care, but for people waiting over twelve weeks for help, some support from AI is better than none. Apps can already offer self-care exercises built on proven cognitive-behavioural techniques, and AI can pick up signs of stress, depression or anxiety in the way people write and type.
How long do people wait for mental health help?
When this was written, in 2024, over one million people were referred to mental health services in the UK each year. The services were so oversubscribed that only around half of them got even one session of workbook-based support, and the typical wait was over twelve weeks. While they wait, many people's mental health gets worse, or they drop out, reluctant to be more of a burden.
Can a machine understand people?
AI already sorts recycling, helps pack and deliver what we order, and drives vehicles. AI assistants schedule meetings, answer customer calls and guide customers through their questions online.
In 1972, Allen Newell and Herbert Simon put forward what was then an outrageous view: ‘Humans, when engaged in problem-solving in the kinds of tasks we have considered, are representable as information processing systems.’ Like a computer, the brain stores and retrieves information, works on it and solves problems with algorithms.
AI now makes art that can surprise us with its beauty, and the fact that we can debate whether it is creative matters for how it can be used in mental health.
What can AI do for mental health?
In 2018, BMJ Leader published a qualitative review of advances in AI in health. In it, Professor Erwin Loh said that ‘AI has now been shown to be as effective as humans in the diagnosis of various medical conditions, and in some cases, more effective’, and that ‘when it comes to predicting suicide attempts, recent research suggests AI is better than human beings.’
Through apps, AI can give people self-care exercises that help them manage their mental health and build resilience, based on clinically reviewed and proven cognitive-behavioural techniques.
Natural language tracking can find patterns in chats, social media posts and emails that may go with stress, depression or anxiety. Some claim it can detect depression from the way someone types on their phone, since depression affects how we move. AI can watch for these signs and prompt a conversation exactly when it is needed.
What does it cost?
It risks dehumanising healthcare. There is no relationship between doctor and patient, and empathy and trust, central to any treatment, are human things. AI can make art, hold a conversation, guide and help, but can it relate to a person's emotional state? Can empathy be coded at all?
Is any help right now better than none?
We think so, absolutely. Mental health care is broader than the doctor's sofa of the movies, and sometimes simply having somewhere to unburden your thoughts makes a great difference.
Poor mental health is estimated to cost the global economy $2.5 trillion a year in lost productivity, and with services stretched to breaking point, help is harder to reach. Anything that helps is valuable, and AI's command of language has taken giant leaps.
For a more definite answer, in 2024 the University of Plymouth was trialling an AI chatbot app, measuring its effect against a control group of people on waiting lists. Technology will keep developing to help us, not only with tasks but with our wellbeing and happiness. How 4 Roads builds intelligent self-service.
Updated on 19 September 2026.
