Artificial intelligence (AI) could revolutionize the delivery of mental health care, helping to streamline clinician workflows and assist with diagnostic and treatment decisions. Yet, before AI can be integrated into practice, it is necessary to understand perspectives of these tools to inform facilitators and barriers to their uptake. We gathered data on clinician and community participant perspectives of incorporating AI in the clinical management of eating disorders. A survey was distributed internationally to clinicians (n = 116) with experience in eating disorder treatment (psychologists, psychiatrists, etc.) and community participants (n = 155) who reported occurrence of eating disorder behaviors. 59% of clinicians reported use of AI systems (most commonly ChatGPT) for professional reasons, compared to 18% of community participants using them for help-related purposes. While more than half of clinicians (58%) and community participants (53%) were open for AI to help support them, fewer were enthusiastic about their integration (40% and 27%, respectively) and believed that they would significantly improve client outcomes (28% and 13%, respectively). Nine in 10 agreed that AI may be improperly used if individuals are not adequately trained, and could pose new data privacy and security concerns. Most agreed that AI will be convenient, beneficial for administrative tasks, and an avenue for continuous support, but will never outperform human clinicians on relational skills. While many clinicians and community participants are open to the use of AI in eating disorder treatment and recognize its possible wide-ranging benefits, most remain cautious and uncertain about its implementation.
Read full abstract