Professor Katie Holt holds the Alieen Dunham Chair in History at The College of Wooster.
I’ve long been an advocate for teaching with Wikipedia as a pedagogical approach to help students strengthen their information literacy skills. The guided trainings provided by Wiki Education help students think about how to critically evaluate different sources of information. Our class discussions and student reflection push them to consider how and why some topics get more coverage than others, identify crucial content gaps, and do the research and writing to make Wikipedia more representative. Plus, for first year students in particular, writing for Wikipedia is an effective hands-on way to practice using citations to show sources of information.
Inspired by my College of Wooster Ed Tech colleague Dr. Jon Breitenbucher’s “Guidelines for AI in Higher Education: Practical, Student-Centered Course Design Tips & Resources”, this fall I wondered if I could take a similar approach to helping students think critically about information they get from generative AI queries. Up until now, a lot of my conversations with students about AI have focused on potential violations of academic integrity if they were to present ideas that were not their own, or talking through student concerns about the environmental costs of AI. My hope was that through drawing on the information literacy approach I use in teaching with Wikipedia, my students and I could interrogate generative AI queries as a source of information about the past. How accurate were the answers these models supplied? What were the sources LLM relied on? Did students see any biases in the way AI outputs framed their explanations of events?
My goal in this approach was to get students to think critically about what is gained and lost when they use generative AI as part of their research strategies. If students will be asked to use AI in future workplaces, it is important that they understand potential pitfalls and the need to not take output at face value. I also wanted students to think through the problematic nature of the lack of transparency about source material for many LLMs. There is a serious potential for generative AI responses to amplify inaccuracies and biases found in their source material.
I developed this assignment for a small workshop I teach that focuses on different forms of public facing scholarly communication for Juniors starting on their mentored research projects. In this workshop, students practice communication strategies for different genres and audiences, including oral presentations, TikTok style public history videos, Instagram posters, and writing for Wikipedia.
Early in the semester, students are doing preliminary research on their topics. For our unit on writing for Wikipedia, my assignment had two parts: first, a comparative analysis of Wikipedia as a source of historical information and second, the identification of a content gap that they would address by adding two new high-quality sources. In terms of their history communication, I emphasized these edits as a way to integrate scholarly perspectives and practice writing in Wikipedia’s neutral, accessible tone.
I incorporated Wiki Education’s new Using generative AI with Wikipedia training module as part of my students’ training and found it very helpful. This new module orients students to their responsibilities as they contribute new content to Wikipedia to incorporate information from high-quality, reliable sources. In particular, it explains why editors should not use generative AI to produce new Wikipedia content because they cannot verify the information provided and because of the pervasive use of Wikipedia as sources for LLMs.
For part 1 of this assignment, I asked students to critique background information about their research topics found in three different types of sources: a traditional encyclopedia article, a Wikipedia article, and a generative AI query. How do the interpretations, the sources used, and the reliability of the facts presented vary in different places where people start their quests for historical information?
I did allow students who said they had ethical objections to using AI for any purpose to instead choose a third reference source for comparison. About one fourth of my students chose this option. Many of us assume that our students are enthusiastic adopters about this new technology, but I think it is important to remember that there is a wide range of student stances regarding generative AI.
Interestingly, many students were frustrated by their research with traditional encyclopedia articles. I wasn’t expecting this response, and think it was a reflection of students’ expectation that they would be able to find reference articles on their exact topic rather than a broader contextual entry.
In our discussions, one theme that came through was students’ newfound appreciation for Wikipedia as a starting place for information. Many of my students had been forbidden to use Wikipedia as a source in high school, so had not been looking at its articles through a scholarly lens. They praised Wikipedia’s easily verifiable sources and links to more information – although they noted a great deal of inconsistency in both coverage depth and citation use. Still, most of students found Wikipedia articles that provided a good historical overview of their subjects written in accessible language and with helpful suggestions of outside links and starting academic sources.
For the AI queries, student praises focused on ease of use, convenience, speed, and the accessible format of the output. One student wrote that Generative AI “does have a nice, pleasing structure of bullet points compared to the other encyclopedia articles.” Another student using generative AI for the first time remarked: “I was surprised by just how much information it was able to provide, and how familiar and personal the tone seemed…[this assignment] did give me an interesting perspective on why people use (or abuse) it in the ways that they do.” This was a helpful reminder for students about how important it is to adjust format, tone, and approach for their intended audience in their own writing.
Students also appreciated that with personalized AI queries, they could do a more narrow, focused probe about one aspect of their topic. Some thought this could be a helpful first step as they brainstormed for topic ideas or essay structure: “…for general knowledge that is proved by many sources, generative AI can help begin research and brainstorm. Since generative AI pulls from multiple sources from the internet, I would not use it for research rather organization or brainstorming.” Again, I thought that students’ discussion of where AI queries could be helpful or provide a starting point for their own critical thinking was useful.
However, my students were very critical of the quality and reliability of historical information generative AI provided. Students found their AI-generated responses contained inaccurate or misleading information about key events, in some cases supported by links that led to nothing. Some information was oversimplified, smoothing over the past and presenting events as less complex or contested than they were. Students missed useful information about scholarly debates over how to interpret the past incorporated in discipline-specific traditional encyclopedias.
A final point of concern was what students saw as the AI models’ regurgitation of internet information without citations. Several students speculated they could recognize the role of Wikipedia as LLM source material: “What I thought was interesting when comparing these was that the AI overview was nearly an exact (although shortened) version of Wikipedia. I knew that AI models took from things but I never realized how direct it was or how clearly it made up sources.” Students came away with a deeper understanding of the potential for AI to magnify content gaps and bias.
I’m very pleased with how this approach worked in getting students to think critically about some of the most common sources of historical information. As one student concluded, “Overall, I have learned that traditional encyclopedias, Wikipedia, and AI work better as tools or mediums for further research than as ends in themselves.” This isn’t surprising given their end goals of producing a substantial academic research project.
Students came away from this project with a newfound sense of the importance of improving freely accessible information about historical topics available for the general public. They found writing for Wikipedia to be a meaningful way to use their skills of historical analysis and access to high-quality information through our library to make a significant improvement in Wikipedia’s coverage of the past.
Interested in incorporating a Wikipedia assignment into your course? Visit teach.wikiedu.org to learn more about the free resources, digital tools, and staff support that Wiki Education offers to postsecondary instructors in the United States and Canada. Apply by December 1 for priority consideration for spring 2026 courses.
