PURPOSE
Artificial Intelligence (AI) and Large Language Models (LLMs) are
profoundly impacting medical education. This study used a
"learner-as-developer" approach to enhance microbiology,
pharmacology, and clinical medicine integration via team based
student-driven, AI-enhanced study tools on CNS infections, fostering critical
AI appraisal.
METHODS
One hundred ninety-one second-year medical students participated. They
completed Pre- and Post-Assignment Quizzes and Surveys. Students worked in 33
teams to generate integrated study tools (questions/flashcards) on CNS
infections using prescribed AI models. A Post-Assignment Quiz and Survey
assessed knowledge gain and self-reported perceptions of the AI tools.
RESULTS
One hundred ninety-one students completed tool generation. Analysis of quiz
scores revealed a highly significant knowledge gain (p<0.001), with a mean
increase of 3.24 points. Pre-survey data showed 89% used AI for academic
purposes. Post-assignment, 82.7% rated the tools as highly or moderately
efficient. Trustworthiness concerns were notable: 78.6% perceived tools as
moderately/slightly trustworthy, with only 10.2% rating them highly
trustworthy.
CONCLUSION
AI tools are highly effective as productivity assistants for low-leverage
tasks like generating practice questions, particularly in integrating
microbiology with clinical medicine. Learners-as-Developers reported high
satisfaction with efficiency and a strong intention for future use. However,
limited perceived trustworthiness suggests AI serves as an auxiliary tool
rather than a core learning source for experienced learners. To drive deeper
learning impact, pedagogical strategies must prioritize training in advanced
prompt engineering and implementing transparent content verification to
mitigate user-perceived risk of superficial information.