Number
803
Name
Development and Testing of a Research Compliance Google Gem for Student-Focused Research Project Development
Date & Time
Sunday, June 7, 2026, 5:30 PM - 7:00 PM
Location Name
Oglethorpe Ballroom
Authors
Jared Vineyard, Idaho College of Osteopathic Medicine Mirando Robledo, Idaho College of Osteopathic Medicine Joanna Owen, Idaho College of Osteopathic Medicine Michael Wells, Idaho College of Osteopathic Medicine
Presentation Topic(s)
Technology and Innovation
Description
PURPOSE: In support of the growing demand for research experiences in the
UME setting, conducting analysis projects using information from large
health-related databases presents a cost-effective and enriching opportunity.
We sought to establish a process and procedure for connecting faculty and
students at our institution to these online resources. The guidelines for
working with each resource vary, and the technical expertise required
suggested that a scaffolded pathway would best reduce barriers to project
design and launch in a compliant manner.
METHODS: Through literature and website review, we have begun to outline a
pathway for researchers with this interest. To promote compliance within
proposal development and database use, we further sought to employ an
AI-based tool that is easily accessible in real-time. Google Gemini is the
preferred platform at our institution, so this took the form of a Google Gem.
We populated this internally facing, content-limited Gem with our research
compliance documents and standard operating procedures using a
retrieval-augmented approach, enabling the tool to reference these materials
during use. We challenged the efficacy of this tool through hypothetical
inquiries and refined inclusion of source documents as needed to address
deficiencies.
RESULTS: Iterative testing has identified multiple deficiencies in the
Gem’s responses. Addition of more source documents targeting these topics
partially ameliorates these issues. The breadth, depth, and rigor of our
inquiries was high, with the intent to simulate a multitude of student and
faculty inquiries related to project design, execution, resource use, and
compliance protocols. The refinement process continues, and we will report
updated findings.
CONCLUSION: Recent advances in AI-based tools reveal optimistic promise for
the support of faculty and student research projects at all stages. Training
of such tools on institutional policies and procedures can minimize
unintended and noncompliant uses of highly regulated resources and open new
avenues for research.