The role of evidence-based medicine in health technology assessment in the era of artificial intelligence

Purpose: The rapid expansion of artificial intelligence (AI) in clinical practice has introduced new methodological and institutional challenges for health technology assessment (HTA). This editorial discusses how evidence-based medicine (EBM) can be applied and adapted to the assessment of AI-enabled medical technologies, with particular attention to implications for the Korean HTA system.Current concepts: AI-enabled medical technologies, including image analysis, digital pathology, risk prediction, clinical decision support, and generative AI applications, differ from conventional interventions in several important respects. Their performance depends heavily on the characteristics and biases of the training data, may change after deployment through continuous learning, and can be difficult to interpret because of model opacity. Their clinical value also depends on how clinicians and patients interact with AI systems in real-world care settings. International reporting and appraisal frameworks, including CONSORT-AI, SPIRIT-AI, TRIPOD+AI, and DECIDE-AI, provide useful tools for evaluating validity, applicability, and risk of bias. However, post-deployment monitoring based on real-world evidence and periodic reassessment is needed to detect performance degradation, algorithmic drift, and inequities across patient subgroups.Discussion and conclusion: The core principles of EBM—transparent evidence synthesis, structured assessment of certainty, and evaluation of patient-centered outcomes—remain essential in the AI era. These principles should be implemented through a life-cycle approach that links regulatory approval, HTA, reimbursement decisions, evidence generation, and reassessment. In Korea, the National Evidence-based Healthcare Collaborating Agency can play a central role in developing AI-specific assessment methods, supporting evidence generation, and strengthening data infrastructure to promote the safe, effective, equitable, and value-based use of AI-enabled medical technologies.