Clinicians and consumers have long been interested in using purpose-built chatbots to provide mental health support. Specifically designed therapy chatbots are now available direct-to-consumer, even though researchers have yet to establish their efficacy, safety and viability. However, whatever their clinical merits or limitations, the role for specialised therapy chatbots has been overshadowed by the increasing number of people using AI companions and general-purpose generative AI for mental health support. Reports have implicated these offerings in instances of user self-harm, prompting calls for more robust regulation across the entire field. This Element examines the opportunities, risks and legal landscape of AI for direct-to-consumer mental health support and considers a response of distributed regulatory networks. This approach abandons any pretence of a single body of law providing an effective and palatable response for concerns raised by therapy chatbots and the challenges posed by evolving technologies operating in sensitive domains.
