Describing the Framework for AI Tool Assessment in Mental Health and Applying It to a Generative AI Obsessive-Compulsive Disorder Platform: Tutorial.

Abstract

As artificial intelligence (AI) technologies occupy a bigger role in psychiatric and psychological care and become the object of increased research attention, industry investment, and public scrutiny, tools for evaluating their clinical, ethical, and user-centricity standards have become essential. In this paper, we first review the history of rating systems used to evaluate AI mental health interventions. We then describe the recently introduced Framework for AI Tool Assessment in Mental Health (FAITA-Mental Health), whose scoring system allows users to grade AI mental health platforms on key domains, including credibility, user experience, crisis management, user agency, health equity, and transparency. Finally, we demonstrate the use of FAITA-Mental Health scale by systematically applying it to OCD Coach, a generative AI tool readily available on the ChatGPT store and designed to help manage the symptoms of obsessive-compulsive disorder. The results offer insights into the utility and limitations of FAITA-Mental Health when applied to "real-world" generative AI platforms in the mental health space, suggesting that the framework effectively identifies key strengths and gaps in AI-driven mental health tools, particularly in areas such as credibility, user experience, and acute crisis management. The results also highlight the need for stringent standards to guide AI integration into mental health care in a manner that is not only effective but also safe and protective of the users' rights and welfare.