Generative AI for Academic Study
What this handout is about
You’ve likely heard of AI tools such as ChatGPT, Microsoft Copilot, Perplexity, or others by now. These tools fall under a broad, encompassing term called generative AI that describes technology that can create new text, images, sounds, video, etc. based on information and examples drawn from the internet. In this tip sheet, we will focus on potential uses and pitfalls of generative AI tools for studying.
If you’re curious more generally about what generative AI is, how it works, and what role prompting plays in the process, you can look at the modules in Carolina AI Literacy, the UNC Libraries’ Student Guide to AI Literacy, and the Writing Center’s handout on “Generative AI in Academic Writing.”
Before we begin: Stay tuned to your instructor
Instructors’ opinions on the use of AI tools may vary dramatically from one class to the next, so don’t assume that all your instructors will think alike on this topic. Consult each syllabus for guidance or requirements related to the use of AI tools. If you have questions about if/how/when it may be appropriate to use generative AI for studying, be sure to seek input from your instructor.
Note that when your instructors authorize the use of generative AI tools, they will likely assume that these tools may help you think, write, or do calculations—not think, write, or do entire calculations for you. Keep that principle in mind whenever you use AI for study purposes of whatever kind. You can maintain your academic integrity and employ the tools with the same high ethical standards that you apply to every area of your academic work.
Potential uses
Exploring Topics through Questions
Generative AI may help you unpack and master material from class in ways that are more conducive to your own learning style and prior understanding. You can prompt AI tools to do the following:
- Explain individual concepts and theories, as well as the connections between a set of them
- Apply concepts to specific subjects, tasks, or types of problems
- Generate comparisons and contrasts about a given concept or its application (“How is X different from or similar to Y”?)
- Reformulate a topic with a series of images and metaphors
As you experiment with prompting, pay close attention to the output, and revise your prompt accordingly. Prompt engineering is a key piece of using AI effectively. For example, if you’ve started with a general prompt to compare and contrast choosing between a Physics and English major, and the result isn’t satisfactory to you, try being more specific. You could say, “List five similarities and differences between choosing a Physics major and an English major, and discuss the pros and cons of each major.” You could take it one step further and give the AI tool a persona: “Discuss the pros and cons of each major from the point of view of a very practical grandparent.”
AI tools can apply more abstract scientific theories to real-world situations and problems and provide concrete examples. In other words, generative AI can help you explore different practical applications for the concepts and theories you are studying. This, in turn, can be very helpful when preparing for quizzes and exams that emphasize application over mere memorization. Don’t be afraid to ask generative AI tools a variety of questions to explore a topic from different angles.
Generating Study Materials
AI may be useful in generating different kinds of study materials like practice exams and problems, quizzes, study guides, or flashcards. For example, many calculus courses at UNC use an AI-powered tool called Edfinity to assign students homework online. As a student in these classes, you can use this platform to generate “Practice Similar” problems that allow you to practice on random variations of assigned homework problems. AI may also be useful in creating study materials for humanities courses. For example, you could ask AI to make a timeline of the plot of a work of literature to help you remember the details. These materials could then be used as the basis for conversations and study sessions with your professor, TA, a group of classmates, or friends. Research has shown that self-testing–or making your own practice tests based on the course materials you need to cover–is one of the best ways to prepare for an exam, and in general leads to higher GPAs. AI can help with this, but keep in mind that for such tasks, it is important that you make the prompt as specific as possible.
Uses in STEM
Besides creating practice problems, you can also use generative AI tools to generate step-by-step solutions to some types of math problems, and explain the relevant formulas and foundational concepts behind them. If you are a STEM student, AI tools may help you to grow your math skills by understanding the logic behind a solution and practice interrelated problems of comparable length and difficulty. Absent other resources or the immediate feedback of a professor/TA, quick access to practice problems can help you make your own practice tests in preparation for exams.
You can also benefit from the use of generative AI even when chatbots “hallucinate” incorrect answers. By checking the AI output for relevance and accuracy, you can learn not just how to solve problems, but how to assess the validity of your answer(s), an important metacognitive skill. In many cases, AI will produce multistep output, part of which is correct and part of which may be plain wrong. So, by isolating the incorrect responses from the correct ones, and being able to explain why certain steps are “more correct” than others, you can fill gaps in your knowledge and learn how to set up solutions to problems faster than you could have if you started them from scratch. In other cases, AI may claim multiple solutions to a given problem, so learning how to assess the correctness of several potential solutions will encourage you to seek a more nuanced understanding of the problem.
Summarizing Texts and Concepts
You can put longer, more difficult texts into an AI tool and ask for a summary of the key points after you are done reading. AI summaries may also serve as a guide to orient you to the text before you read. After reading the summary, you can read the full text to analyze how the author has shaped the argument, to get the important details, and to capture crucial points that the AI response may have omitted from the summary. The same can be done with concepts, theories, and individual ideas that you want to have summarized so you can break them down and apply them to your given study task or course. You may also want to compare the outputs with other summaries such as found on Wikipedia or library sources to cross-reference and expand your understanding.
Managing and Planning Studying
One of the most important parts of any successful plan of study is managing your time wisely. Generative AI tools can help you develop a weekly or daily schedule and study plan to do so. Depending on the specificity of your prompt and the information you input–for example, when and where you want to study on a given day and/or the specific content you need to study–AI may be useful in helping you come up with a detailed plan of action to tackle your test or maintain a constant study routine.
Potential pitfalls
Barriers to Active Learning and Metacognition
You can use AI tools in such a way as to block active learning and metacognition. While AI can be a shortcut that provides ready answers and information with little to no effort on your part, you do not want to skip over the necessary process of independent learning and undermine your own education. Using AI tools in some instances may help you submit your homework on time, but if you don’t learn the material on your own, you may be caught short when test time comes. It is important to consider whether or not your use of AI is actually helping or hindering your development as a learner. Consider how you will check your mastery of the material–without AI-–as you study for a test. Assessing your depth of understanding is especially important any time you will need to apply knowledge. You want to transcend memorization or summary of information to prepare for most classes.
False or Inaccurate Information
AI tools derive their responses by reassembling language in their data sets, most of which has been culled from the internet. As you learned long ago, not everything you read on the internet is true, so it follows that not everything reassembled from the internet is true either. Beware of clearly written but factually inaccurate responses from AI tools. They can produce information that may seem plausible to you, but is in fact partly or entirely fictional. This may become a problem for AI responses that include citations. While the citations may seem reasonable and look correctly formatted, they may be incorrect or just not exist at all. For example, AI tools may invent an author, produce a fake book title, or incorrectly attribute language to an author who didn’t write the quote or wrote something quite different. As noted above, AI hallucination continues to be a real danger. You should always double-check source or citation AI outputs. You can use the library to follow up on suggested source material to learn more when legitimate citations seem promising.
Biased Responses
Again, AI tools are drawing on vast swaths of language from their data sets–and everything and anything that has been recorded in them. Accordingly, the tools mimic and repeat distortions in ideas on any topic in which bias easily enters in. You should consider and be on the lookout for biases in responses generated by AI tools. This is one of the more complicated and easily overlooked pitfalls to AI-generated information.
Cost and Data Privacy
Use of generative AI raises concerns about cost and access. While many AI tools remain free and open-access, these versions may not be as robust or accurate as newer iterations that require monthly subscription fees. The use of generative AI also comes with data-privacy risks. What will these technologies do with the information you put into their systems? Will your ideas become part of their property and profit? Be aware of how AI tools will use your information. Check the privacy policies and terms of use of any AI platforms that you experiment with or adopt.
Issues with Computational Mathematics and Writing Code
Two areas in which current AI tools have difficulties are computational mathematics, and, more generally, any kind of programming, coding, or quantitative calculation. In some cases, generative AI like ChatGPT and WolframAlpha can solve advanced math problems in areas such as geometry, trigonometry, calculus, and differential equations with an acceptable degree of accuracy. But, quite often, these solutions are not the most efficient. Moreover, AI tools struggle to solve more complex math problems with the same degree of accuracy, especially when these problems demand particular knowledge from the real world that is beyond the “database” at the time the output is produced. At present, humans are still much better at logical reasoning than AI tools like ChatGPT.
In addition, AI tools like ChatGPT are less precise than scientific calculators for performing basic computations. ChatGPT has been trained on a vast text dataset, so it responds to prompts based on context rather than facts, which is an unreliable way to handle numbers. AI tools attempt to combine the arithmetic of traditional calculators with the articulation of natural-language processing, resulting in a “calculator” that is prone to fabrication.
If you use an AI for calculations, be sure to check its output. What’s more, while many AI systems are capable of generating tons of practice problems for a given math topic–like functions that can be differentiated in calculus–beware of the fact that adaptively-generated practice problems can show little variance in difficulty and scope, making it hard for you to scaffold your understanding as intended. Practicing many closely-similar problems may convince you that you know more than you actually know, a phenomenon known as the “Fluency Illusion.” Know that you may be required to verify and reconstruct any calculation or computation generated through AI that you use in your assignments. You don’t want to risk relying on false data as you study, so make sure you are using AI mindfully in a way that will help you study.
A final note
Acquainting yourself with these AI tools may be important as your thinking and study skills grow. While generative AI is new and still under development, it may prove useful for you to understand in your current academic life and in your career after you leave the university. Beginning to experiment with and develop an understanding of AI tools at this stage may serve you well along the way.
Additional Resources
- Carolina AI Literacy Modules
- UNC Libraries’ Student Guide to AI Literacy
- Writing Center’s handout on “Generative AI in Academic Writing”
Note: This tip sheet was updated in September 2024. Generative AI technology is evolving quickly. We will update the document as the technology and university landscapes change.
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivs 4.0 License.
You may reproduce it for non-commercial use if you use the entire handout and attribute the source: The Learning Center, University of North Carolina at Chapel Hill
If you enjoy using our handouts, we appreciate contributions of acknowledgement.