The use of generative artificial intelligence as a mental health resource: a rapid review on prevalence, conceptual boundaries, and public health implications

Generative AI is already being used as a mental health resource by a considerable proportion of the population.Among the most robust estimates, 16% of US adults and 25% of adolescents in England and Wales reported using genAI for explicit mental health advice or support during the preceding 12 months.Recent nationally representative surveys suggest even more widespread use: 48% of people aged 16-39 in Germany reported discussing general psychological issues with genAI and 51% of those aged 18-34 in Czechia reported using these tools for mental health or emotional problems within preceding 30 days.These numbers also illustrate a key problem in the evidence base: studies measure different behaviors under the broad label of mental health-related use.To improve interpretation and future measurement, this review distinguishes three layers: i) companionship and social interaction, ii) emotional support and self-regulation, and iii) explicit mental health advice and support.Substantial methodological differences remain even within these layers, and reliable population estimates from low and middle income countries are largely absent.Better standardized and repeated monitoring is needed to understand the scale and development of this rapidly evolving public health phenomenon.