Examinining the AI Scholarly Research Writing Landscape
Brandon Patterson, Technology Engagement Librarian at the Spencer S. Eccles Health Sciences Library, is leading an innovative research project to explore how AI-powered tools can support scholarly writing and undergraduate research. The initiative, titled“ Examining the AI Scholarly Research Writing Landscape,” focuses on integrating practical AI applications within teaching sessions that the library supports, while also providing an opportunity to discuss the limitations of using AI in a research context.
Project Goals
The study set out to:
- Determine which AI scholarly research tools are reliable and accurate, and how they align with current information practices used by trained librarians.
- Identify how these tools might fit into workflows developed by librarians who teach scholarly writing.
- Consider education and training needs, as well as deployment strategies, to encourage adoption if the tools prove helpful.
Key Findings
After examining a dozen distinct AI-powered tools, Patterson and his team found that Elicit, Scite, and Undermind were rated more highly than others. However, the research also revealed a critical limitation: a lack of transparency and reproducibility in the results these tools provide.
Because of these concerns, Patterson recommends that such tools be used during the exploratory phase of research development, while traditional scholarly inquiry methods remain essential for maintaining high-quality reviews and scholarship.
Why It Matters
This project underscores the importance of AI literacy and critical evaluation of evidence in scholarly work. The libraries have formed a working group to strategize ways to assist
Brandon recently presented alongside Psychology Professor Jared Branch and students, Nicholas Hauschild and Carlos Flores, at the University of Utah's AI Summit.
For more information, you can connect with Brandon by emailing b.patterson@utah.edu.