Determinants of AI Use and Related Concerns In Social Work: A Comparative Study of Japan, South Korea, and the United States

Background: The use of artificial intelligence (AI) in social work is rapidly expanding across both formal and informal contexts. Although views on AI are mixed, social work practice continues to evolve in response to technological change. This study examines social workers’ concerns regarding AI and their determinants, with particular attention to technostress, professional identity, and practice orientation. While AI can improve efficiency and convenience, it may also substitute aspects of core professional tasks, raising questions about professional roles and identity. Moreover, digital technologies may enhance relationship-building while potentially weakening client–worker relationships. This study adopts a comparative perspective across Japan, South Korea, and the United States, where AI use in social work is increasingly relevant.

Methods: A cross-sectional online survey was conducted among social workers in Japan, South Korea, and the United States in April 2026, in cooperation with professional associations and local institutions. Approximately 900 responses were obtained across the three countries. The questionnaire assessed demographic and organizational characteristics, AI use in practice, organizational adoption and support systems, AI-related concerns, professional identity, and relational practice orientation. Data were analyzed using descriptive statistics and regression-based approaches.

Findings: The proportion of organizations that had introduced AI tools was 34.2% in Japan and 40.2% in South Korea, indicating slightly higher adoption in Korea, while the U.S. showed comparable levels of organizational exposure to AI. Frequent use was reported by 33.3% in Japan and 61.9% in South Korea, suggesting substantially higher utilization in South Korea, while the U.S. indicated moderate levels of regular use. Across contexts, AI was most commonly used for document-related tasks, information search, training, and idea generation to support planning. Concerns about AI were widely shared, particularly regarding potential errors, bias, and privacy issues, and were more pronounced in Korea than in Japan.

Multivariate analyses in Japan and South Korea indicated that technostress, professional identity, and relational practice orientation were significant predictors of AI-related concerns, with higher technostress and stronger professional and relational orientations associated with greater concern. In the U.S., professional identity was positively associated with AI-related concerns, whereas technostress was not, suggesting that professional values may play a central role in shaping these concerns across contexts. Greater AI use may heighten concerns among practitioners who prioritize relational, client-centered practice, raising questions about its impact on roles and client relationships.

Conclusion and Implications: AI use in social work has expanded beyond document support to more practice-oriented domains, such as support planning and information gathering. While its benefits are recognized, concerns about risks remain strong. The findings suggest that increased AI use does not necessarily reduce concerns; rather, its diffusion may be accompanied by heightened awareness of ethical and professional risks. Across contexts, AI-related concerns appear to be shaped less by technological burden than by professional identity and relational commitments. Promoting AI use therefore requires not only technical training but also clear guidelines and a clearer articulation of AI’s role within human-centered practice, ensuring alignment with professional values.