AI-PACE: A Framework for Integrating AI into Medical Education

The integration of artificial intelligence (AI) into healthcare is accelerating, yet medical education has not kept pace with these technological advancements. This paper synthesizes current knowledge on AI in medical education through a comprehensive analysis of the literature, identifying key competencies, curricular approaches, and implementation strategies. The aim is highlighting the critical need for structured AI education across the medical learning continuum and offer a framework for curriculum development. The findings presented suggest that effective AI education requires longitudinal integration throughout medical training, interdisciplinary collaboration, and balanced attention to both technical fundamentals and clinical applications. This paper serves as a foundation for medical educators seeking to prepare future physicians for an AI-enhanced healthcare environment.

💡 Research Summary

The paper “AI‑PACE: A Framework for Integrating AI into Medical Education” addresses the widening gap between the rapid adoption of artificial intelligence in clinical practice and the comparatively slow evolution of medical curricula. The authors performed a systematic literature search across PubMed, MEDLINE, ERIC, and the Allen AI (Asta) platform for articles published between 2016 and 2025, using keywords related to AI, medical education, competency frameworks, and physician training. From an initial pool of 643 records, 23 papers met the inclusion criteria: (1) proposal or evaluation of a structured AI curriculum, (2) target audience of undergraduate medical education (UME), graduate medical education (GME), or continuing medical education (CME), and (3) English language. Purely technical papers, commentaries without curricular proposals, and non‑U.S. studies were excluded.

A thematic analysis of these 23 papers identified three recurring shortcomings in existing models: (a) fragmentation into short‑term bootcamps or elective workshops, leading to rapid knowledge decay; (b) a specialty‑centric focus (radiology, ophthalmology, etc.) that leaves a “generalist gap” for primary‑care physicians; and (c) a near‑absence of the affective domain—trust calibration, automation bias awareness, and preservation of empathy in AI‑mediated encounters.

To fill these gaps, the authors propose the AI‑PACE framework, an acronym for Psychomotor, Affective, Cognitive, and Embedded. The first three pillars map directly onto Bloom’s taxonomy domains, while “Embedded” adds a structural dimension ensuring longitudinal integration across the entire training continuum. The framework is visualized as a spiral curriculum: foundational AI literacy is introduced in the first year of medical school (cognitive and affective basics), reinforced during clerkships (psychomotor skills and workflow integration), and expanded in residency and CME through specialty‑specific applications and leadership training.

Psychomotor focuses on hands‑on competencies: navigating AI tools within clinical workflows, interpreting probabilistic outputs, performing critical appraisal of algorithmic performance (sensitivity, specificity, AUC), and communicating AI‑derived recommendations to patients. Affective addresses attitudes and values: appropriate trust calibration to avoid both over‑reliance (automation bias) and undue skepticism (algorithm aversion), maintaining patient‑centered care, fostering interdisciplinary collaboration with data scientists, and committing to lifelong learning as AI evolves. Cognitive provides the knowledge base: fundamentals of machine learning, deep learning, health data science, biostatistics, data governance, strengths and limitations of AI systems, and ethical/legal considerations (privacy, bias, regulatory frameworks).

The Embedded pillar operationalizes the framework across three educational stages:

• UME integrates AI concepts into existing modules such as evidence‑based medicine and biostatistics, emphasizing basic theory and ethical awareness.
• GME shifts toward specialty‑specific workflow integration, using case‑based simulations and “human‑in‑the‑loop” verification during rotations.
• CME offers targeted workshops on emerging technologies, institutional AI implementation strategies, and leadership in AI governance.

Recognizing the “faculty gap,” the authors recommend parallel upskilling pathways for junior faculty, ensuring educators can model critical evaluation and entrustment behaviors alongside learners.

The paper concludes that AI‑PACE offers a unified, generalist‑friendly model that overcomes the fragmentation, specialty bias, and affective neglect of prior curricula. Limitations include the English‑U.S. centric literature base and the lack of empirical pilot data to validate the framework’s impact. Future work should involve multi‑institutional pilots, longitudinal outcome assessments, and mechanisms for rapid curriculum updating in response to AI advances.

In sum, AI‑PACE provides a comprehensive, longitudinal, and domain‑balanced roadmap for preparing physicians to work safely, ethically, and effectively with AI tools, thereby aligning medical education with the realities of an AI‑enhanced healthcare ecosystem.