Dr. Laura Hardin Marshall, Webster University
Dr. Carolyn I. Brown, Webster University
With Assistance from Writing Center Coaches (Spring 2025)
A recent study by Tully, Longoni, and Appel found that when people have lower AI literacy (when they don’t have as much knowledge about what AI is or is capable of), they are “more likely to perceive AI as magical and experience feelings of awe” about what AI can do and are therefore more receptive to it (2). They hypothesize that “efforts to demystify AI may […] reduce its appeal.” Here are a few basic details and strategies for AI that may help “demystify” these tools and promote more critical understanding of AI and its uses.
Find the why: Instructors expect students to engage in intellectual work for a reason. By doing the activity, what specific skill is developed? What knowledge is acquired? How might that skill or knowledge be useful in other situations (academically or professionally)? And how might AI fall short? Students who do not see the purpose of an assignment (e.g., perceive it as “busy work”) will likely be more inclined to outsource the work.
Help students understand how it works: Generative AI language models are essentially smart predictive text. Just because words are statistically more likely to go together doesn’t mean they make sense or communicate accurate information/ideas. This often leads to major content errors, including false information and fabricated sources. Exercises that analyze and identify flaws in AI output can help take some of the perceived “magic” out of the tool.
Understand prompt engineering: The quality of AI output relies considerably on the quality of the prompt used. Give AI tools specific parameters, roles, and expectations (be careful, though: it won’t always follow parameters because it’s predictive text). For example, what skill level and formality should the output be? What’s the purpose and context, who is the audience, where and how will it be shared? The more information the user can provide about the rhetorical situation and expectations for the final product, the more serviceable the results will be.
Correct/revise the output: Stress to students the reality of effective AI use. Because AI output can have serious flaws, users need to do a lot of work verifying the information (including the sources used) to make sure that the content is appropriate. Users are responsible for the quality and accuracy of the content and how they use it. Also consider: It takes a lot of time and effort to research and verify AI content (and rewriting can be more time intensive than writing). Could that time be better spent by learning the information and doing the work right the first time?
Ineffective uses of AI: Using AI output uncritically (without learning or attention to the information generated) and without intensive correction (using it “as is”) is generally a bad idea. To avoid simply copying and pasting, writers may be tempted to outsource the heavier parts of the work but still “be involved” in easier parts of the writing process, e.g., use AI to process source information and produce the majority of the paper, with the writer only engaging in light polishing. In general, AI output will likely be lacking in personal, individualized insights and critical thinking, so even though the writer contributed to some of the process (the end stage), it won’t resolve the major issues unless they put a lot of work into correcting and revising the content.
Potentially effective uses of AI
Brainstorming: Using AI in the early stages of a project is becoming increasingly accepted in academic settings. It can be useful for generating lists of potential discussion points, finding keywords or search terms, identifying areas of controversy or differing perspectives, etc.
Planning and strategy: AI tools can offer a roadmap, including steps and to-do lists for how to accomplish the work. It can also help the writer evaluate potential paths to take and predict the results of using one strategy over another.
Organizing and structure: AI tools can be helpful for suggesting ways to organize or (re)arrange information as well as indicating conventional expectations for formatting and structure. It can also suggest potential reasons for structural choices, which can be useful for transitions.
Mechanics and polishing: AI generally does a serviceable job in producing Standard Academic English; it can also make serviceable adjustments to phrasing and clarity. If using AI tools this way, though, be sure to carefully read the revised output; AI is not always going to “understand” the meaning, so its paraphrasing can sometimes change the message or intent.
Borderline uses of AI: The ways writers can “effectively” use AI vary greatly: course by course, instructor by instructor, assignment by assignment. A good general rule is that when it’s time to produce the bulk of the content (the product that will be turned in for a grade), that’s typically a good time to set it aside. For example, treat AI output like we do Wikipedia—it can be good to learn from it, to use it as a starting point, but we don’t typically want to use it as a source. But are there times and places we would want to use AI/Wikipedia? Probably! The following are some ways that writers could use AI, but they should also ask themselves, “Should I use AI this way?” Consider what is gained or lost by letting AI do the work before deciding.
Summaries and synopses: AI can provide a summary of source information, which can help writers quickly process and understand information. However, be cautious: Did the AI tool summarize the work accurately? It’s more effective to use AI to supplement reading or to verify understanding. Also, do we have the right to share an author’s work with AI tools? If the source is open access, we can safely assume that AI tools have/will mine the source, but it’s highly unethical to upload copyrighted information into AI tools without the author’s permission.
Letting AI be a teacher: Writers often accept AI output without taking time to verify or even really read the content. If AI has produced most of the work, be prepared to go through the paper and make a list of the main ideas or significant talking points. Take the time to process what AI has said about those ideas, and then reflect on whether the information conflicts with the user’s own knowledge/opinions or ideas discussed in class, etc. Learn from what AI has “said” and be able to talk intelligently about those ideas.
Paraphrasing and polishing: Students often ask whether it’s okay if they provide the ideas and have AI “polish” them, e.g., input their informal writing (e.g., notes or freewriting) and let AI turn it into more formal writing. If formal writing expectations are part of the course goals, this would likely not be an ethical use of AI. However, if written communication and writing conventions are not a learning objective and the goal is the student’s understanding of the content and ability to generate ideas, this may be an effective way to use AI.
Common Reasons for Using AI + Effective Responses/Alternatives
There are many reasons why writers might potentially use AI, even if it’s not allowed by their instructor. When having discussions about AI use, it’s helpful to explore each student’s personal reasons for using the tool and (if they’re receptive to changing their methods) offering realistic alternative strategies that can directly address their concerns.
Want a higher grade: Point out the flaws in most AI writing; the AI product may be more “polished,” but the content/information likely will not meet the assignment objectives, resulting in a lower grade instead of a higher one (including the risk of a potential integrity infraction).
Not confident in their own abilities: Offer reassurance. In many cases, human-written ideas are much more interesting than what AI can produce. Additionally, most instructors care more about seeing where students are, ability-wise, so that they can offer guidance and support to promote learning and growth. There are also plenty of great resources to develop skills, including the Writing Center!
Unsure of how to start/effective writing practices: Share tips, tricks, and resources. For example, writing is typically easier when it’s broken into stages. What are the exact steps to follow (chances are the Writing Center has a handout for it!)? It also helps to approach writing in smaller chunks: instead of thinking, “I have to write a 7-page paper,” think, “I’m writing one paragraph about X.” Take it a paragraph (or even a sentence) at a time! If getting started seems intimidating, write in a more informal style first; getting ideas down on paper/screen (regardless of quality or polish) is what’s important.
Short on time or intimidated by the research process: Use time-effective reading strategies (read the abstract, introduction, and conclusion, then skim section headings and body paragraphs). Pick only 3-4 essential sources and focus on those. Sometimes, the perfect source simply doesn’t exist, so find what research is tangentially relevant and use critical thinking to apply it to the topic.
Need help with polishing: There are better tools for proofreading and polishing! Most students have access to Word through their school accounts, and the Editor (grammar and spellcheck) feature generally does a good job with polishing (especially if the options are personalized in the settings).
Lack of knowledge/interest in the topic: Ideally, pick a different topic, one that the writer is more familiar with. If that’s not an option, use AI (or, even better, a librarian!) for a starting list of keywords. Use those as search terms for preliminary database research. Skim/read the abstracts of several relevant articles to gather a baseline of information; expand and organize the keywords into an outline.
Unsure of what to say: Take one interesting section from a source (a paragraph, sentence, chart, graph, etc.) and take a few moments to think about it. Simply react to what the source is saying and write down those thoughts (freewriting)! This makes a great starting point for a discussion.
Writer’s block or lack of ideas: Brainstorm with a writing coach, a classmate, or a friend. Sometimes, even just speaking ideas aloud (or talking to a pet or an inanimate object) can be just as useful.
References
Tully, S., Longoni, C., and Appel, G. (2025). Lower artificial intelligence literacy predicts greater AI receptivity. Journal of Marketing, 0(ja). https://doi.org/10.1177/