Applying Knowledge
Jody Ondich
Creating Labels and Descriptions
AI can be used to generate a description of a selected subject, create a summary or label based on established conventions in the field, and locate additional resources to deepen their understanding. This approach integrates research, analysis, and creative interpretation with responsible AI use.
Example: Generate a Label and Summation of an Art Piece
Goal
The goal of this assignment was to use both art and archaeology to address the stories about King David of the United Kingdom of Israel. Accurate descriptions were important, as was structuring an art label for the painting involved. Various prompts elicited materials that could be added to, restructured, or adapted for use in creating a project.
How It Was Done
To successfully complete this project, students must engage with one or more works of art alongside references to relevant archaeological evidence that supports or contextualizes the topic. In the example provided, the focus is on the narratives surrounding King David of the United Kingdom of Israel. While this example includes a pre-selected artwork, students could alternatively use AI tools to identify appropriate artistic representations related to the topic.
Initial AI prompts included a request for a general description of Caravaggio’s David with the Head of Goliath, a list of standard art labeling conventions, and a completed label for the painting using those conventions. A subsequent set of prompts sought information on the Tel Dan Stele, an archaeological artifact referencing the “House of David.” Students were expected to evaluate the AI-generated content for the credibility of sources and the accuracy and depth of detail. This approach provides a foundation for constructing a project that explores both the historical and biblical dimensions of King David through interdisciplinary inquiry.
Results
Despite refining the prompts, Copilot proved limited in its ability to locate relevant images and was unable to provide information on the carbon dating of materials from the Tel Dan archaeological site—information that was readily accessible through a basic internet search. In contrast, Copilot performed well in generating a description of the painting and producing a draft of the corresponding art label using appropriate conventions.
While the AI-generated materials were usable as a starting point, they required clarification and revision. The quality of sources varied significantly, ranging from reputable museum and educational websites to less academically acceptable sources such as Wikipedia and generic Bible study platforms. As a demonstration of what AI can contribute to the research and drafting process, the results were valuable; however, as a polished final product, they fell short of academic standards.
Potential Issues
- One limitation encountered was Copilot’s inability to locate an image of the Tel Dan inscription, despite its easy availability through a standard Google image search. This highlights the need to supplement AI tools with traditional search methods, particularly for visual materials.
- Evaluating the quality of sources generated by Copilot is essential. In this instance, the tool relied on a mix of sources, including less academically rigorous platforms such as Wikipedia and general Bible study websites, alongside more credible museum and education-based resources. This variability underscores the importance of critical source assessment when using AI tools for academic work.
Best Tools for the Job
- Working with images is likely going to work better in another AI, such as Dall-e or even just checking Wikipedia or Google advanced searches. Getting thorough and thoughtful labels and descriptions from Copilot, however, was easy. When one copies and pastes the responses from Copilot into a document, Copilot even creates footnotes and a list of sources from the copied materials.
About the author
name: Jody Ondich
institution: Lake Superior College
Jody teaches philosophy at Lake Superior College in Duluth, and is part of a faculty development team there. Her focus areas are OER development, digital accessibility, AI use, and online course review programs. She has written 3 OER texts, housed at the University of Minnesota Open Textbook network.
