Page 149 - AIH-2-4

Artificial Intelligence in Health AI-driven personalized learning in residency

Figure 2. Control chart for time spent on the AI-powered learning platform
Abbreviations: AI: Artificial intelligence; LCL: Lower control limit; UCL: Upper control limit.

Figure 3. Total duration per resident on the AI-powered learning platform, expressed as mean and SEM, with the outliers removed
Abbreviations: AI: Artificial intelligence; SEM: Standard error of the mean.

governance around content curation is essential for scaling While this study provides valuable insights into
these tools responsibly. 21 AI-enhanced medical education, several limitations
In the present study, we posit that the ICE played a must be acknowledged. First, the educational research
critical role in upholding ethical standards, mitigating study was conducted within a single internal medicine
misinformation, and ensuring the pedagogical validity residency program, without a comparison or control
of AI-generated educational content. By integrating group, potentially limiting the generalizability of findings
automated safeguards such as source verification, toxicity across different specialties and training environments.
filtering, and bias mitigation, ICE directly addresses Second, although engagement and quiz performance were
concerns raised in prior literature regarding the reliability strongly correlated, long-term educational outcomes, such
of AI-driven instructional material. 22,23 This technological as board examination performance or clinical decision-
advancement represents an ethically sound, meaningful making improvements, were not assessed. In addition,
evolution over earlier AI-enhanced learning systems resident perceptions of AI-based learning were collected
by combining adaptive personalization with structured through surveys. However, the study did not include
content governance. In addition, early AI technologies qualitative interviews or focus group discussions, which
such as Verbot™ demonstrated the potential for digital could have provided more profound insights into learner’s
agents to enhance classroom engagement and delivery. experiences and preferences. Recent studies have shown
24
As such, the platform supports individualized learning that generative AI tools can support learner’s motivation
trajectories and reinforces the academic integrity necessary and improve knowledge retention in higher education
for implementation in medical education. 25 contexts. Future research should explore multi-
26

Volume 2 Issue 4 (2025) 143 doi: 10.36922/AIH025130023