Physician role transformation in AI-driven healthcare: Role demands, adverse work outcomes and AI competencies as resources

Extant research has sparingly tackled the theme of emerging role transformation of physicians in an AI-driven healthcare environment, and the requisite AI competencies for sustaining their job performance. This research draws on the Job Demands–Resources (JD-R) framework as a sensitising analytical lens to conceptualise physician role transformation as changes in core work characteristics that generate job demands, the resulting adverse work outcomes as part of the health impairment pathway, and AI-related competencies as job resources that can buffer these impacts. While JD-R also encompasses a motivational pathway, this study deliberately centres on strain-related mechanisms considering the early and evolving stage of AI implementation in healthcare. We empirically explore these relationships through a theory-informed inductive qualitative designcomprising thirty-five physician interviews across various specialties and healthcare settings across eleven countries. Based upon systematic thematic analysis, we identified, five emergent role transformation dimensions of physicians, namely, ‘ AI Integrator’, ‘Data Steward’, ‘Workflow Architect’, ‘Ethical Guardian’, and ‘Reflective Practitioner’, and five adverse outcomes: “ Emotional Strain”, “Cognitive Fatigue’, ‘Performance Impairment’, ‘Role Uncertainty’, and ‘Social Role Conflict’. We further theorise five domain-specific AI competencies as buffering job resources , namely, ‘ AI Fluency’, ‘AI Clinical Competence’, ‘AI Risk Mitigation Competence’, ‘Responsible AI Leadership Competence’, and ‘ AI Relational Competence’. By operationalising role transformation and competence development within the JD-R framework, this study extends job design theory into AI-intensive healthcare contexts. Practically, the identified themes offer actionable guidance for healthcare workforce development, organisational support, and capability-building to ensure that physicians are prepared to navigate the new role and are well equipped in an increasingly AI-driven healthcare environment. • Conceptualises AI-driven physician role transformation as structured job demands within the JD–R framework. • Identifies five adverse work outcomes arising from AI-related work redesign via the health-impairment pathway. • Develops a taxonomy of five physician AI competencies as demand-specific job resources. • Empirically grounded in 35 physician interviews across eleven countries. • Extends JD–R theory by integrating AI-enabled work design with demand–resource matching logic.