Incorporating GenAI in Assignments and Activities
Students and instructors continue to explore the possibilities and limits of using generative AI in classroom assignments. Beyond ways GenAI can support students’ content knowledge (for example, through GenAI tutors or AI-generated self-examination), GenAI tools can also be applied in both writing-to-learn contexts and major writing assignments. This resource will outline some principles for incorporating GenAI into elements of students’ writing process and identify both potential benefits and risks of various forms of GenAI intervention.
Important Preliminaries: Course Context Considerations
With any pedagogical innovation, instructors often balance competing values, priorities, and practical constraints. Practices that work effectively for a class of 20 students may not scale easily to a class of 80. Scaffolded assignments are great, but they may feel confining for very advanced students or overwhelming for students juggling multiple jobs or family obligations. No teaching technique is guaranteed to work for all students, so variation and experimentation can help address more students’ needs overall. When considering inclusion of GenAI activities in class, consider these four important questions:
- What do you want students to know or be able to do by the time they complete the course?
- What is your specific teaching context (class size, student population, mode, curriculum, etc.)?
- What prior learning experiences do your students have in your field, or what range of experience might students bring to the course?
- Where and how do you imagine that students will incorporate GenAI into their coursework?
Learning outcomes can guide you in selecting an appropriate role for GenAI, whereby students maintain the most important cognitive labor in the course rather than shifting it to a GenAI tool. Generally, the closer students are to novice-level proficiency, the less helpful GenAI interventions may be.
During the first week of classes, exercises like pretesting and surveys can help instructors better gauge what students have retained from prior learning experiences and how much they have incorporated GenAI into their study and writing habits.
A Recommended Protocol for Incorporating GenAI in Writing Assignments
- Students should do some reading and writing first, with GenAI used to supplement or augment their initial human effort.
Regardless of how they incorporate GenAI in their coursework, encourage students to read and write something themselves first, such that their own thinking and labor initiate every course activity. This can mean precluding students from prompting GenAI with inappropriate instructions (e.g., generate four topics for a research paper on climate-based migration) and instead prompting the GenAI with questions related to their own initial research (e.g., ask open-ended questions that will help me narrow down my suggested topics to a feasible research question for a ten-page paper). Evidence from multiple sources suggests that starting with GenAI tools rather than independent thinking can have negative consequences, including reduced cognitive effort, a false sense of mastery or comprehension, and anchoring effects that limit students' creativity and lateral thinking. Similarly, using a GenAI summary as a check for understanding (to confirm students' comprehension of reading material) is preferable to relying exclusively on a GenAI summary that scrubs detail and specificity in the name of simplicity.
- Any instance of GenAI use should include predicting generated outcomes and their limitations.
Early research in students' use of GenAI raises serious concerns regarding students' judgment of the accuracy and credibility of GenAI-produced content. When students have less background knowledge or expertise, their assessments of credibility are strongly influenced by peripheral factors, including style, tone, and arrangement. The declarative and authoritative tone of GenAI output, and the speed at which it is delivered, can imbue a false sense of confidence in its quality1. Before students incorporate a GenAI tool in their writing process, ask them to address, in their own words, some important clarifying questions:
- What do you hope GenAI output will provide, and how is your prompt designed to get that result?
- What do you suspect the output will look like? (e.g., Will it contain references to research literature and citations, or just links to sites? Real examples or hypotheticals?)
- How will the GenAI output help me to make better use of my own resources?
- How will I fact-check any assertions made by the GenAI tool?
These preliminary questions cue students to recognize that output quality is strongly influenced by prompt design and the features of training sets/models, and encourage skepticism toward results.
- Require students to document their assessment of any GenAI output, even if they decide not to incorporate it in their research and writing process.
Just as prediction can encourage students to take a more skeptical view of GenAI outputs, making assessment of GenAI output an explicit process step can also encourage students to be more careful in considering the value of GenAI contributions. Often, the authoritative tone of a GenAI output can persuade students that the output they received is both accurate in content and sufficient in scope (despite the fig-leaf warnings on most GenAI tools about the risk of hallucination). Creating an explicit step for the assessment of output quality and coverage can encourage the habits of fact-checking that are a crucial element of GenAI literacy and writing more broadly.
Reflection and Assessment Questions for Different GenAI Uses
| Writing Process Stage | Human-First GenAI Use | Human Review and Assessment |
|---|---|---|
| Brainstorming | The student freewrites or concept maps, then prompts GenAI to coach: Read my initial freewrite and ask me three open-ended questions to prompt further thinking. |
|
| Research | The student creates a list of research questions, then prompts the GenAI with a specific question: Suggest keywords I should use in academic database searches to find genomics research related to these research questions. |
|
| Reading | The student reads and summarizes an article, then generates a GenAI summary to check their work. |
|
| Data Collection | The student writes a survey draft, then prompts the GenAI: Review this survey draft and suggest plain language edits that would make it more accessible to an audience of first-year, traditional-age college students. |
|
| Data Analysis | The student organizes their data and identifies an appropriate statistical test, but asks GenAI to generate R code to run that test: Suggest the steps and R code to run a MANCOVA on the attached data. Independent variable: test_condition, dependent variable: test_score, covariates: student_gpa, hours_studied The student selects an appropriate data visualization type and prompts GenAI for directions to use the software: Tell me the steps to create a stacked bar chart in Excel, change the colors, and add data labels. |
|
| Drafting | The student prompts the GenAI to respond as a tutor: You are a tutor who helps students with college writing. Engage me in a conversation with lots of non-leading questions that will help me assess whether this outline is clear, potentially make revisions, and get motivated to start writing. |
|
| Revision | The student prompts the GenAI to respond as a reader: You are a patient newly diagnosed with diabetes. You have only a high school education. Read this draft pamphlet, and tell me about your experience reading it. What is clear and what is unclear? How would you respond to it? The student assesses the response and revises it based on their intended impact on the audience. |
|
| Editing | The student prompts GenAI to help them improve their editing skills: The student uses GenAI to generate practice opportunities: Create a sample text with common grammatical errors. I’ll correct it and ask you to check my work. |
|
| Citation | The student writes citations and prompts GenAI to check for consistency with a style guide. (Better option: use a citation generator, which also saves work without introducing the risks of GenAI tools in a very specific process.) |
|
| Translating | The student generates a translation of a section of their text—Provide a translation of this paragraph while maintaining meaning and tone as closely as possible—and manually checks it for accuracy. |
|
| Coding | Students in an advanced course outsource simple steps they can already perform confidently, but complete high-level tasks themselves. |
|
Many faculty members and instructors who are investigating GenAI-augmented assignments are generously sharing their activities in collections like Text Gen Ed: Teaching with Text Generation Technologies. Many of the assignments included are geared toward developing students' GenAI literacy and tool familiarity, but many also address critical questions about the ethical and appropriate application of GenAI technologies. Similarly, many publishers are promoting their own built-in GenAI tools to supplement instruction.
An important word of caution: assignments designed for one context may not be appropriate for all contexts, so always ensure that your assignments and activities are compliant with the UMN GenAI policy:
- Only recommend the use of UMN-approved GenAI tools and platforms.
- Do not require students to purchase subscriptions or pay fees to use other platforms, even if they are connected to published course materials.
- Caution students to review data privacy and copyright rules associated with the materials they use.
Can I require students to use GenAI tools?
Course policies differ on GenAI use. In some cases, the course or curriculum requires students to use and explore GenAI tools: Computer Science, Information Technology Infrastructure, and Carlson’s Business Analytics and GenAI curriculum are easy examples. If you intend to require all students to use GenAI tools, it’s wise to disclose this to students up front.
Some students and scholarly organizations argue that students have a “right to refuse” to use GenAI tools. The University of Minnesota has no explicit policy on students' right to refuse to use GenAI, and it is unclear whether refusing to engage with GenAI should be understood as a reasonable content accommodation based on strongly held religious beliefs or practices (which can require content accommodation in other contexts). In general, if a student indicates a committed objection to the use of a GenAI tool, it’s possible to complete the assignment (or a reasonable alternative) in a way that does not require the use of GenAI, and such a change will not diminish the student's ability to meet a learning goal, providing a GenAI-free alternative is an inclusive practice.
1As GenAI tools allow users to select among lower resource models (often called fast models) and more advanced, resource-demanding models, this trend may persist. Because so-called ‘advanced reasoning models’ may generate text through a process that implies human-like research, students may be even more susceptible to attributing credibility to GenAI tools even though they lack human capacity to understand meaning and value.