Clinician Perceptions of AI in Care: Cross-Sectional Survey

AI is increasingly being explored as a tool to enhance efficiency, access, and diagnostic accuracy in mental health care. However, the perspectives of clinicians, who are central to the delivery and oversight of care, on the use of AI remain underexamined.

This study aimed to explore clinicians' perceptions of AI in clinical care, including perceived benefits, risks, barriers to implementation, and training needs.

A cross-sectional survey was distributed (via email and social media) from August 2024 to November 2024 to mental health professionals in the United States. The survey aimed to capture attitudes toward AI integration, perceived utility, and adoption-related challenges and facilitators in mental health care. Quantitative data were analyzed using descriptive statistics, whereas open-ended responses were analyzed thematically to identify key insights.

Respondents (N=62) were mostly (43/62, 69.4%) not currently using AI in practice. The most frequently endorsed benefits included reduction in time spent in coding and billing (32/60, 53.3%), automated intake (23/61, 37.7%), and improved documentation (23/61, 37.7%) by mental health professionals. However, most expressed discomfort with using AI in patient care, and concerns were raised about inaccurate outputs (47/61, 77%), algorithmic bias (41/61, 67.2%), reduced accessibility (31/61, 50.8%), and job stability (22/61, 36.1%). Barriers to adoption included clinician resistance (46/61, 75.4%), lack of validation (41/61, 67.2%), and challenges with technical integration (26/61, 42.6%). Most respondents (54/61, 88.5%) believed that specialized training in AI ethics and applications was moderately, very, or extremely important for clinicians. Open-ended responses reinforced concerns about dehumanization, cultural insensitivity, ethical accountability, and insufficient technological literacy.

Mental health professionals in this sample viewed AI as a potentially useful adjunct to care but not a replacement. Ethical concerns, limited trust, and a strong emphasis on the human dimensions of therapy suggest that implementation must proceed with caution. Clinician-informed strategies, ethical frameworks, and targeted training are essential to support the responsible and effective integration of AI into mental health practice.