Adopt a differentiated integration strategy for GenAI in undergraduate medical education rather than uniform adoption¶

Strategy · All strategies
Evidence · 5 claims (4 for, 1 mixed) · 1 study (1 quant-synthesis), q3 · 1 of 1 report an effect size · 5 claims rest on one study

Description¶

The review concludes that GenAI's educational benefits are concentrated in specific domains and recommends matching adoption to those domains. It states that GenAI's "efficacy is concentrated in cognitive clinical skills, academic writing, and nursing education", and that "these findings support a differentiated integration strategy rather than uniform adoption". The strategy directs educators to deploy GenAI where pooled evidence shows benefit and to await stronger evidence elsewhere.

Design Implications¶

Context¶

Requirements¶

• Future high-quality, multi-center, large-sample, and long-term RCTs to robustly confirm effects, as the article states

Constraints¶

• Evidence quality for primary outcomes was downgraded to low, and other outcomes to very low, per the GRADE assessment

Target Learners¶

• Undergraduate medical students, including clinical medicine and nursing majors

Target Learning Goals¶

• Basic knowledge acquisition
• Clinical skills
• Academic writing proficiency

Claims¶

• Nursing students show significant, homogeneous clinical skills benefit from GenAI (SMD = 0.98), and the overall clinical-skills effect differs between majors [~W]
• GenAI significantly improves academic writing scores for both clinical medicine and nursing students [+W]
• Nursing students show significant knowledge improvement with GenAI (SMD = 1.11) unlike clinical medicine students [+W]
• GenAI does not significantly improve clinical skills of clinical medicine students compared with traditional teaching [+W]
• GenAI does not significantly improve knowledge exam scores of clinical medicine students compared with traditional teaching, with high heterogeneity [+W]

Related Strategies¶

• Prioritize autonomous motivation and scaffold concept-focused AI support tailored to learner trajectories rather than providing uniform access alone
• Differentiate T2I pedagogy by student profile and adopt a stage-sensitive approach to AI literacy and assessment in studio courses
• Frame LLM integration as an assessable AI literacy competency with differentiated integrity policies and reasoning-focused assessment redesign
• Differentiated, profile-matched university AI-readiness and career-education support instead of uniform AI-literacy programs
• Position immersive technologies and virtual patients as supplementary components within blended learning rather than stand-alone replacements for established clinical communication training

Examples¶

-

Key Sources¶

• Xie Guanli, Liao Jianglong, Tang Xiaoxia, Fu Han, Liu Duo, Deng Li, Wang Tao, Hong Xin. (2026). The impact of generative artificial intelligence on clinical skills and knowledge acquisition in medical undergraduates: a systematic review and meta-analysis. BMC Medical Education. https://doi.org/10.1186/s12909-026-09458-3