Abstract: Objective To explore effective training pathways improving teachers′ cognition of artificial intelligence(AI) and rapidly enhance their ability to apply AI in medical education. Methods A baseline questionnaire was conducted among teachers at the School of Basic Medical Sciences of Peking Union Medical College to assess their current understanding of AI-enabled medical education, acceptance, skill level and training needs. Based on the analysis, a cyclic enhancement training model (“Cognition Guidance-Skill Instruction- Output Motivation- Practice Validation”) was developed. A follow-up questionnaire was used to comprehensively assess the training outcomes. It encompassed basic information, training participation, effectiveness and satisfaction evaluation,and outcome assessment. Results The proportion of teachers with insufficient AI understanding dropped from 66.66% to 29.27%. Eighty-five percent rated the training as highly or very targeted and effective; over 90% reported clear theoretical explanations and strong practical inspiration. Overall satisfaction (satisfied/very satisfied) was 90.24%, 80.49% confirmed a significant improvement in AI application competencies, and 97.56% expressed expectation for future training. Conclusions The cyclic enhancement training model effectively improves teachers′ cognition of AI integration into medical education and rapidly enhances their proficiency in applying AI tools to teaching practices.

Key words: artificial intelligence, medical education, training, forced output method, learn by doing