Biopsychosocial care prioritisation by health professionals and the public after watching AI-generated patient videos, and perceptions of AI-mediated communication in obesity and mental health care

Background: Artificial intelligence (AI) is increasingly used in health communication,yet evidence on health professionals’ (HPs) and public perceptions remains limited.
Objectives: To compare how HPs and public participants prioritised care needs afterwatching AI-generated patient video case studies and to examine broader perceptionsof the potential impact of AI-mediated communication, such as chatbots, virtual agentsand avatar-led information, on healthcare interactions in obesity and mental healthcare.
Design: A cross-sectional online survey with an embedded video case studyassessing care prioritisation and separate general items assessing perceived AI mediatedcommunication impact on healthcare interactions.
Methods: A total of 276 participants (93 HPs; 183 public participants) wererandomised to one of four pre-recorded, avatar-based AI-generated video casestudies simulating a consultation with a patient experiencing obesity and psychologicaldistress. Participants ranked biopsychosocial care priorities and digital lifestyle contentand reported perceptions of AI’s impact on healthcare communication. Group (HPsand public) differences were analysed using Mann–Whitney U and chi-square tests.
Results: Psychological and emotional support and lifestyle change were consistentlyprioritised, while pharmaceutical and surgical interventions ranked lowest. HPsprioritised cultural beliefs significantly more than public participants (p=0.016), andloneliness/social support more highly (p=0.007), while public participants gavesignificantly greater importance to gut health (p<0.001). AI impact perceptions weredistributed across “help”, “hinder” and “no opinion”, with no cohort association (χ²(2,N=276)=0.432, p=0.806). Loss of human touch was the most frequently selectedcommunication-related concern (HPs 55%; public 48%).Conclusion: HPs and public participants prioritised psychological and lifestyledomains over medical interventions. AI-mediated communication was viewed ambivalently, with trust-related concerns centred on loss of human touch,miscommunication, reliability and human oversight.Practice implications: AI-supported health communication should complementrelational care, preserve empathy and personal connection, maintain human oversight,and communicate transparently about its role and limitations. Co-design should guidedevelopment and implementation.