This content is drawn from a report authored by the AU Library's Artificial Intelligence Experimentation Working Group. You can read the groups full report covering experiments with the use of AI in library work and recommendations to library leadership in the American University Research Archive.
This working group adopted a project-based approach to AI experimentation, based on the recommendations of Gupta & Gupta (2023). All experimentation centered around streamlining specific projects or workflows, and the tools chosen were based on their applicability to those projects. Such an approach prevented the subgroups from focusing too heavily on any single tool and narrowed the focus of the experimentation to direct ways to enhance operations.
The AI Experimentation Working Group was formed in Spring 2024. Anyone who was interested in experimenting with and researching AI was invited to join the working group, participation did not require prior AI experience. Based on survey responses, the Working Group created four subgroups focused on the following specific project themes:
The subgroups identified specific projects and AI tools for testing, with the object of streamlining or improving these tasks and workflows.
Throughout the summer of 2024, subgroups engaged in in-depth experimentation of the application of AI to the identified projects. Each subgroup documented their prompts, successes, and challenges using a standardized template. This documentation was used to summarize findings for a showcase where subgroups shared their results with the library, and to inform this recommendation report for library leadership based on their initial testing and feedback.
The experimentation of this working group focused on the strengths and limitations of the tools as applied to each library task. Issues surrounding cost, equity of access, learning curves, and user privacy and/or copyright concerns for each tool were also addressed.
Research questions include:
Gupta, V., and Gupta, C. (2023). Leveraging AI technologies in libraries through experimentation-driven frameworks. Internet References Services Quarterly, 27(4), 211-222.