Summary
Provides consensus guidance and evaluation methods for adult-facing generative AI wellness applications used for emotional support, general wellness guidance, or coping assistance, addressing product scope, harmful advice, crisis response, over-reliance, privacy, bias, transparency, and monitoring.
Healthcare Implications
Developers and implementers should clearly distinguish wellness support from clinical care, disclose system limitations, test high-risk and crisis scenarios, implement escalation and dependency-reduction safeguards, protect sensitive data, evaluate performance across user groups, and monitor changes after deployment.