AI-Supported Early Detection of Burnout and Depression in Social Work Professionals: Effort-Reward Imbalance, Emotional Labor, and the Differential Mediating Role of Surface Acting

Background. Social work professionals constitute one of the most psychologically vulnerable occupational groups, characterized by sustained exposure to emotionally demanding client interactions, chronic resource scarcity, and structural organizational pressures. Although effort-reward imbalance (ERI) is an established predictor of burnout and depression, the mechanisms through which it produces psychological harm—and the potential of artificial intelligence (AI) to support early detection—remain underexplored. Aim. This study examined surface acting, a form of emotional labor, as a mediator linking ERI to emotional exhaustion and depression, and derived implications for a conceptual AI-based monitoring framework. Method. A cross-sectional online survey was conducted with N = 98 social work professionals in Germany (M age = 36.9 years, SD = 8.6; 68.4% female). Instruments comprised the Effort-Reward Imbalance Scale, the Maslach Burnout Inventory–Emotional Exhaustion subscale, the Patient Health Questionnaire-9, and the Surface Acting subscale of the Emotional Labor Scale. Parallel mediation models were estimated using bias-corrected bootstrapping with 5,000 resamples. Results. ERI strongly predicted both emotional exhaustion (β = .88, p < .001) and depression (β = .87, p < .001). Surface acting partially mediated the ERI–exhaustion pathway (indirect β = .11, 95% CI [.01, .22]) but did not mediate the ERI–depression pathway (indirect β = −.003, 95% CI [−.11, .10]). Conclusion. Emotional labor contributes to the exhaustion pathway but not the depression pathway, indicating that the two outcomes arise through partly distinct mechanisms. This differential pattern carries direct implications for the design of AI-supported early detection systems, which should track the shared resource-depletion trajectory rather than assume a single mediating mechanism.