As the use of generative AI continues to grow, understanding how people are using these tools for mental health has become an important public health question. This narrative review brings together findings from surveys and research published between 2024 and 2026 to estimate the prevalence of AI use for mental health support and examine the challenges in measuring it.

Drawing on available evidence, the authors estimate that around 27% of AI users have used AI for some form of mental health support, including emotional support, companionship, wellbeing guidance, or therapy-like conversations. The review also highlights major gaps in the evidence base, including inconsistent terminology, varying survey methods, and the absence of agreed definitions for AI-assisted mental health support, making more precise estimates difficult.

The paper calls for greater consensus on definitions and research standards to better understand how AI is being used in mental health, while informing policy, regulation, and future service development.

Original publication: npj Digital Public Health (2026). Read the full article