Heterogeneity in AI attitudes, anxiety, and acceptance among psychology students and psychotherapy trainees: a domain-specific latent class analysis

Introduction Artificial intelligence (AI) is increasingly proposed as a support tool in psychotherapy, yet little is known about how psychology students and psychotherapy trainees psychologically orient toward such tools beyond aggregate acceptance scores. Methods This study used latent class analysis (LCA) to explore heterogeneity in psychology students' and psychotherapy trainees' ( N = 286) attitudes toward AI, technology acceptance, readiness, and anxiety in relation to two AI-supported clinical tools: an automated feedback tool and a treatment-recommendation tool. Six separate latent class models were estimated across the domains of general AI attitudes, technology acceptance, readiness, job-related anxiety, learning anxiety, and concerns about AI risk and autonomy. Results Across domains, three- to four-class solutions consistently emerged, indicating that participants cannot be described by a single, uniform orientation toward AI. Across several domains, moderate or mixed response patterns were prominent, whereas other classes reflected more favorable or unfavorable orientations, including skepticism, elevated learning- or job-related anxiety, or heightened concern about AI risk. Because each domain was modeled separately, these results describe domain-specific latent response patterns rather than a single integrated psychological profile spanning acceptance and anxiety simultaneously. Discussion Findings are discussed in relation to the Unified Theory of Acceptance and Use of Technology (UTAUT), technostress, and professional-identity-threat perspectives, and suggest that training for AI-supported psychotherapy tools may need to be tailored to distinct subgroups rather than a uniform trainee population. Given the exploratory, cross-sectional, secondary-data design, findings should be treated as hypothesis-generating rather than confirmatory.