Number
822
Name
DESIGNING AN AI PROMPT WORKSHOP FOR PRECLINICAL MEDICAL AND DENTAL STUDENTS
Date & Time
Sunday, June 7, 2026, 5:30 PM - 7:00 PM
Location Name
Oglethorpe Ballroom
Speakers
Authors
Christopher Tze Yu Wong, A. T. Still University Kirksville College of Osteopathic Medicine
Hermandeep Sandhu, A. T. Still University Kirksville College of Osteopathic Medicine
Matheu Wong, A. T. Still University Kirksville College of Osteopathic Medicine
Madhuran Selvaa, A. T. Still University Kirksville College of Osteopathic Medicine
Yohei Norimatsu, A. T. Still University Kirksville College of Osteopathic Medicine
Bill Miller III, A. T. Still University Kirksville College of Osteopathic Medicine
William Brechue, A. T. Still University Kirksville College of Osteopathic Medicine
Renu Agnihotri, A. T. Still University Kirksville College of Osteopathic Medicine
Presentation Topic(s)
Technology and Innovation
Description
PURPOSE
General artificial intelligence (AI), though not yet firmly established in
medical education, is beginning to influence how students study, retain, and
apply knowledge. Despite this interest, structured guidance on effective and
responsible AI use remains limited. To address this gap, we examined how
preclinical students and faculty across ATSU campuses (KCOM, MOSDOH, SOMA,
and ASDOH) currently use AI in learning and teaching.
METHODS
A survey was distributed to all ATSU students and faculty to assess
experiences, challenges, and perceived needs regarding AI integration. Using
these data, we designed and led a workshop focused on Google Gemini, the AI
platform within ATSU’s secure Google Workspace system. The session
demonstrated strategies for generating COMLEX Level 1 and USMLE Step 1–style
questions, improving retention, identifying content gaps, and using AI as a
complementary study partner. Participants received tools to define Gemini’s
role, outline tasks, and refine prompts for accurate responses. Shivam Vedak,
MD (Stanford University School of Medicine), an expert in AI, opened with a
lecture on general AI use, followed by a demonstration and guided practice.
The workshop was recorded for asynchronous access. After completing it in
their preferred modality, students submitted a post-workshop survey.
RESULTS
Pre-workshop survey data showed that of 232 respondents, 114 (49.1%)
reported AI failure due to poor prompts; of 216 respondents, 134 (62%)
requested training in advanced prompting; and of 219 respondents, 125 (57.1%)
and 108 (49.3%) preferred asynchronous and in-person instruction,
respectively. In the post-workshop survey, nine respondents reported a mean
confidence score of 4.22/5 in prompting and a mean likelihood of continued AI
use of 8.44/10.
CONCLUSIONS
Structured, institutionally supported AI instruction can improve medical
students’ confidence, engagement, and understanding of AI tools. Building on
these findings, we will expand training through additional recorded and
in-person workshops to support innovation in medical education.
General artificial intelligence (AI), though not yet firmly established in
medical education, is beginning to influence how students study, retain, and
apply knowledge. Despite this interest, structured guidance on effective and
responsible AI use remains limited. To address this gap, we examined how
preclinical students and faculty across ATSU campuses (KCOM, MOSDOH, SOMA,
and ASDOH) currently use AI in learning and teaching.
METHODS
A survey was distributed to all ATSU students and faculty to assess
experiences, challenges, and perceived needs regarding AI integration. Using
these data, we designed and led a workshop focused on Google Gemini, the AI
platform within ATSU’s secure Google Workspace system. The session
demonstrated strategies for generating COMLEX Level 1 and USMLE Step 1–style
questions, improving retention, identifying content gaps, and using AI as a
complementary study partner. Participants received tools to define Gemini’s
role, outline tasks, and refine prompts for accurate responses. Shivam Vedak,
MD (Stanford University School of Medicine), an expert in AI, opened with a
lecture on general AI use, followed by a demonstration and guided practice.
The workshop was recorded for asynchronous access. After completing it in
their preferred modality, students submitted a post-workshop survey.
RESULTS
Pre-workshop survey data showed that of 232 respondents, 114 (49.1%)
reported AI failure due to poor prompts; of 216 respondents, 134 (62%)
requested training in advanced prompting; and of 219 respondents, 125 (57.1%)
and 108 (49.3%) preferred asynchronous and in-person instruction,
respectively. In the post-workshop survey, nine respondents reported a mean
confidence score of 4.22/5 in prompting and a mean likelihood of continued AI
use of 8.44/10.
CONCLUSIONS
Structured, institutionally supported AI instruction can improve medical
students’ confidence, engagement, and understanding of AI tools. Building on
these findings, we will expand training through additional recorded and
in-person workshops to support innovation in medical education.
Presentation Tag(s)
Student Presentation