AI and Information Literacy Instruction in the Composition Class and Beyond: Part 1

Prompt engineering and keyword selection are not so different.

A man on a computer screen with a speech bubble above him, probably benefiting from AI and information literacy instruction in the composition class

Ever since OpenAI introduced its large language model chatbot on November 30, 2022, educators, librarians, and librarian-educators have raced to understand ChatGPT and to establish best practices while discouraging plagiarism and other inappropriate uses. 

The challenge is particularly onerous for professors and librarians who work with first-year composition students: a student cohort that must build fundamental research and information literacy skills. Generative AI, like any other tool, can inhibit critical thinking and other vital skills. On the other hand, the emerging technology can be used to reinforce traditional research and citation competencies.

The reality is that most students do (or will) use AI already. Rather than fight it, how can we help them develop the skills that they need to do so effectively and honestly? 

Intersecting interests

Librarians and first-year composition instructors are natural partners in the effort to help students use generative AI in beneficial ways. The composition class, by its nature, requires students to engage in different rhetorical modes while supporting claims with meaningful evidence.

At SUNY Oswego, the institution at which we teach, the College Writing department encourages first-year composition instructors to coordinate with an assigned librarian to conduct a class session in which students are shown how to locate, evaluate, and use appropriate information resources. Professors want students to learn and practice the basics of scholarly research and citation; librarians are the experts in this area. Importantly, first-year composition is one of the few classes that is required for a great majority of students, allowing instructor librarians the opportunity to connect, make a positive first impression, and, one hopes, invite students to return to the library time and again.

Through the course of our in-class collaborations, we have seen how librarians and first-year comp instructors have many intersecting interests. Both of us want students to separate the wheat from the chaff, using well-developed information literacy skills to decide which sources are best to use and in which situations. This, of course, requires students to engage in critical thinking. (Why is Wikipedia okay when double-checking the date ChatGPT went live, but not an okay source to cite in a formal research paper?)

As partners, we reinforce that research and writing are nonlinear processes. An introduction may need to be rewritten for being too boring; a too-specific database search may need multiple attempts before yielding good results. Both research and writing require a great deal of trial and error, particularly for newer students, and sufficient practice to know when one is back on track. 

🌟 Subscribe to the LibTech Insights newsletter for weekly roundups and bonus content, including: 

Similarities between generating AI output & searching library databases

Getting “unstuck” while using AI tools for research is accomplished through judicious prompt engineering, a term that describes how users instruct AI to receive a desired result. This process is strikingly similar to keyword selection for searching databases, a skill that librarians have been teaching for years.

A student may struggle doing a database search for “WWII,” but may find better results when searching for “WWII” AND “economic effects” AND France. Similarly, a student who asks Google’s Bard to “give me ideas for my WWII paper” would be better served by asking, “Can you suggest ten possible argumentative claims related to the economic effects of WWII on France?”

Parallel methodologies

How can partner educators such as first-year composition teachers and librarians best teach students how to approach artificial intelligence in their writing and research? In a September 2023 webinar presented by LibTech Insights, Leo S. Lo, Dean of the University of New Mexico’s College of University Libraries & Learning Sciences, advised that his CLEAR framework is a useful way to get the best results out of AI consultation. The acronym advises users to keep their prompts Concise, Logical, Explicit, Adaptive, and Reflective.  

We have arrived at a similar (and perhaps less scholarly) concept. As unabashed Star Trek nerds, it occurred to us that the characters on the show, through nearly 60 years of adventures, engage in AI prompt engineering on a frequent basis. Of course, no character spent more time talking to the computer than Chief Engineer Geordi LaForge (the character played by librarian favorite LeVar Burton). In particular, the episode of Star Trek: The Next Generation entitled “Booby Trap” demonstrates how students might best approach their experience with ChatGPT or Google’s Bard.

Commander LaForge teaches prompt engineering.

In this episode, the Enterprise-D is caught in a trap set by a long-gone alien civilization. The ancient ship is sapping our heroes’ engines and Geordi must solve the problem. He enlists the help of a hologrammatic representation of Dr. Leah Brahms, the engineer most responsible for creating the Galaxy class warp drive. Through the course of the episode, Commander LaForge acts much in the same manner as our students do during research and writing. He asks the computer for basic facts. He asks to see engine schematics. In effect, he comes up with a string of commands similar to both keywords (for searching databases) and prompts (for engaging with ChatGPT). Taking it a step further, the computer tells him when his theories are likely incorrect and Geordi often argues back. He even asks the computer to speak to him with a different personality. The Star Trek writers obviously take their simulation further than current reality allows, but Geordi’s human-computer interaction is not too far removed from AI’s, and to a lesser extent, library databases’ current capabilities.

Most importantly, Geordi uses all of the assistance from the computer and its hologram to augment his vast knowledge of science and his ship to save the day. The Enterprise’s computer is much the same as ChatGPT: both are essentially chatbots. In addition to the CLEAR framework, perhaps librarians and first-year composition instructors should urge students to think of AI as a second voice, as an advisor. 

At least for now, computers far surpass humans with respect to memory and processing speed, but we analog beings are the only source of true creative and innovative thought. Generative AI can make suggestions and remove an errant comma, library databases can return relevant (and irrelevant) resources, but we want our students to do what they’ve always done best: use the tools around them to hone their thoughts, research, and prose.

In the next installment, we’ll discuss some of the AI-related pitfalls librarians and first-year composition instructors face and offer concrete suggestions that can help our students use the technology in a positive manner.


This piece was adapted from a presentation the authors did at the SUNY Council on Writing’s October 2023 conference: “Writing, Thinking, and Learning with AI: Exploring Relationships of Rhetoric and Artificial Intelligence.”

🔥 Sign up for LibTech Insights (LTI) new post notifications and updates.

✍️ Interested in contributing to LTI? Send an email to Deb V. at Choice with your topic idea.