Facts label for transparent communication of AI Risks in mental health technology
TL;DR
A standardized facts label for AI-DMHTs is proposed designed to enhance transparency and awareness about these tools and their risks to users, patients, and clinicians and serves as a foundation for continued multidisciplinary collaboration regarding the development and governance of AI risk communication in the domain of mental health and healthcare more broadly.
Abstract
With interest in the adoption of artificial intelligence (AI)-enabled digital mental health technologies (AI-DMHTs) among the general population ceaselessly escalating, mental health clinicians are obliged to confront questions about their utility and safety for their practice. However, little guidance exists for developers on how to communicate risks to clinicians who may need to evaluate products for individuals with mental health concerns, individuals who are frequently vulnerable to such risks. We propose a standardized facts label for AI-DMHTs designed to enhance transparency and awareness about these tools and their risks to users, patients, and clinicians. This framework was developed by a multidisciplinary team from the American Psychiatric Association Committee on Mental Health Information Technology through iterative expert review and external clinician consultation, drawing upon existing scholarship in risk communication and informed consent, as well as international AI governance frameworks. The resulting facts label framework is composed of 8 sections: key identifying information, intended use, warnings, risks and limitations, model information, clinical evidence, accessibility and usability considerations, and privacy and security. This research represents a practical step toward responsible utilization of AI-DMHTs and aims to serve as a foundation for continued multidisciplinary collaboration regarding the development and governance of AI risk communication in the domain of mental health and healthcare more broadly.