Considerations for Using AI in the Classroom

Created by Laura Schmidli with Molly Harris, Alison Caffrey, Antonella Caloro, Jonathan Klein, Luis Loya, David Macasaet, Erika Schock, and Parker Story.

Published January 5, 2023. Updated December 8, 2023.

In late 2022, ChatGPT made headlines. This and other generative artificial intelligence (GenAI) tools have many in higher education discussing AI’s potential threat to our work and purpose. To read more from late 2022 and early 2023, see References and Further Reading.

Like with any new technology, one approach instructors can take in the classroom is to transparently make use of it. This allows you to move beyond stress over whether or not students are using GenAI, and instead have a voice in when, why, and how you and your class use it. You can even invite students to consider and explore AI with you (Tufekci, 2022). Although this approach may not work or be right for all instructors and courses, this strategy has the advantages of encouraging student participation in a larger societal question and helping students to develop their own ethical grounding and practical understanding of academic integrity related to AI.

Steps to Consider

This is an accordion element with a series of buttons that open and close related content panels.

1. Reflect on threats and opportunities.

People have raised concerns for student academic integrity as well as for the longer term impact on learner motivation, engagement, and knowledge retention. Others in your discipline may also have specific concerns related to knowledge production, values, and more. What resonates for you?

However, there are also potential benefits. GenAI may save time and effort on routine tasks, provide a new perspective on a problem, or generate content that can be analyzed or critiqued in the classroom.

How do you personally weigh the threats and opportunities in your classroom, and for your own work?

2. Consider what information is safe to share.

Before using any tool or technology, instructors and students should review its data retention and privacy policies – e.g., the OpenAI FAQ for ChatGPT. The general recommendation from campus IT is to only enter information into these tools that you are comfortable making public. See Generative AI @ UW-Madison: Use & Policies for more specific guidance.

In a classroom setting, instructors must also consider information you are asking students to share through use of any non-UW tool, including GenAI. Tools and technologies that are not supported by UW may present risks to instructors and students by not meeting requirements for privacy, security, intellectual property, and records retention. See Use of Third-Party Applications for specific considerations and limitations.

3. Choose one or more GenAI tools to try.

OpenAI’s ChatGPT has continued to be freely available, with a more robust version available for subscription purchase, but requires creating an account using an email address and phone number. Google Bard is also freely available, but does require a Google account.

In contrast, Bing Chat is available for free and does not require an account. Other tools that require less personal information are also available.

4. Investigate strengths and weaknesses of GenAI relative to your course and work.

If provided with a prompt or assessment question from your class, how does the GenAI tool perform? Experiment with questions, tasks, and assignments related to your work, study, or teaching. What are GenAI tools good at, and where do they it struggle? With which routine tasks are you comfortable asking for GenAI assistance?

Keep in mind that writing productive prompts for GenAI requires some skill and practice. See OpenAI’s guidance for designing prompts for ChatGPT.

5. Consider revisions to your assignments.

Making your questions more personal, reflective, specific, local, or complex (i.e., requiring higher order thinking) may make it more difficult for students to use AI to adequately complete them. But more importantly, these changes can make your questions more interesting and valuable to your students.

When designing an assignment where students can use a GenAI tool, always provide an alternative for students who don’t want to share their data with a third party tool (i.e., students shouldn’t have to share their data to be successful). Alternatives could include making part of the assignment optional, the instructor engaging with the AI on behalf of students, students using a different AI language model that does not require personal information, or students using another supplemental source (e.g., interview with a person, responses on social media, review article, etc.) in place of GenAI.

6. Identify and communicate opportunities in your course.

Many students are already using GenAI on their own, and this will likely increase as these tools are integrated into more technologies we use regularly. When instructors acknowledge and model GenAI use, we can guide students in considering the capabilities of these tools, surface differing opinions around fairness, help students align use of these tools with their own ethical frameworks, and even help students grapple with anxiety about the future.

Make your policy on GenAI use in your class overall and in specific assignments (if applicable) clear to students from the start of class. Review the ideas below for exploring AI with your students and consider where GenAI might intersect your course.

Share Your Ideas for Using AI

We’d love to add your knowledge to this guide. What plans, ideas, questions, or concerns do you have for using AI in your classroom?

Share Your Idea

Activity Ideas That Explore Capabilities and Limitations of AI

Ask 20 Questions of AI: In small groups, students collaborate to write 20 questions for a text-generating AI about how it works. In a larger group, they consider what the AI’s responses mean for academic integrity, authority, validity, trust, or other important ideas in your course. 

Analyze AI Perspectives: The instructor asks the text-generating AI to respond to a prompt as a specific person – e.g., a historical figure. Students then critique the AI’s response, drawing on their interpretation of the person’s perspective. 

Predict Where AI Excels: Individually students construct one question or prompt on a specific topic that they think text-generating AI can respond to successfully, and another prompt or question they think AI responds to unsuccessfully. In a larger group, students share their work to identify characteristics of prompts to which AI struggles to respond. For a related lesson outline, see Critical Analysis Across AI Tools & Stereoptypes from the AI Pedagogy Project at Harvard’s metaLAB.

Compare Exam Questions: Individually students write one multiple choice exam question and ask text-generating AI to write a second. In a larger group, students analyze which submitted questions are AI-written, which are human-written, and evaluate which provide a better assessment of learning.

Evaluate AI Output: The instructor uses AI to generate work, like a thesis, short analytical paper, theater dialogue, computer code, image, or even musical composition. In groups students analyze the sample work created by AI, with particular attention to evidence, sources, perceived bias, or other important elements for your course. Students can then revise it for improvement in groups, and share back revisions for comparison. For an example lesson outline, see Correct a Bad Essay from the AI Pedagogy Project at Harvard’s metaLAB.

Activity Ideas That Integrate AI into the Writing Process

Rubric Calibration with AI Writing: The instructor uses AI to generate an essay, thesis, or other written work. Groups then use a rubric to evaluate the AI’s work, and suggest changes or improvement to the rubric. This can help students think about how they define high-quality work, and how a rubric might help identify AI-generated work. 

Planning and Evaluating AI Use: Students create a plan for using AI within a specific assignment, like a scaffolded research paper, where they articulate for which steps it is valuable and appropriate to use AI, and when original thought and creativity are needed and why. This activity works best when sequenced with other activities that explore AI capabilities.

AI Drafting Process: Students use AI to generate a draft of a simple writing assignment. Students then analyze the AI’s writing, focusing on accuracy, bias, or other characteristics important in your course. You may also ask students to improve the AI’s draft to complete a second draft.

 

AI Feedback: Students ask AI to evaluate an initial draft of a short writing assignment, asking AI to focus on a specific element. Students then incorporate any valuable feedback into their work, and share their revisions with a small group. This can help students get another perspective on writing quickly, while encouraging them to consider that feedback critically.

AI Thesis Revision: Students use AI to quickly generate thesis statements on a variety of topics. Individually students revise these statements and share two that are strongest to a group for feedback, including what prompts they provided the AI and what revisions they made to each statement. Students use AI to “get started” by quickly generating ideas, and then leverage their own skills to refine and improve.

Writing with Images: Students or instructors use image-generating AI as part of a reflective writing, freewriting, or creative writing process in any language. Students can use AI to quickly visualize descriptions from their writing, or students or instructors can use AI to generate images that prompt elaboration in writing. For an example lesson outline for generating images, see AI Image Remixing from the AI Pedagogy Project at Harvard’s metaLAB.

Ways to Establish Norms and Boundaries for Using AI in Your Classroom

Co-create Class Expectations: Students collaborate to suggest “ground rules” for using AI within a specific course or assignment. As a class, consider specific tasks where AI is helpful, tasks where AI is unhelpful, when students and instructors think using AI is cheating, and if and how students should cite AI in their work. Ultimately the class will produce documented expectations with the instructor for using AI. For an example lesson outline related to co-creating and discussing expectations, see Debating the Ethics of Generative AI from the AI Pedagogy Project at Harvard’s metaLAB.

Structure an AI Discussion: Based on the “structured academic controversy” model, students are assigned a stance related to using AI in the classroom. Groups collaborate to develop arguments for one perspective and take turns presenting while the other team listens actively and then summarizes the arguments they heard. Individually or in a large group, students then reflect on what they learned, how their personal stance may differ from the stance they were assigned, and whether or not their personal position has changed.

Make Instructor Expectations Clear: Whatever your aims for using (or prohibiting) the use of AI within your classroom, make your expectations clear to students at the semester’s start. This may include adding a statement to your existing academic integrity statement, and reviewing expectations with students in class. Students should also have the option to not use any tool that requires personal information shared with a third party not contracted by the University.

Additional Resources

Connect with Help

We are happy to connect with instructors about using AI in your classroom, designing assignments, or brainstorming solutions for your course. Meetings typically last 45 minutes and take place virtually. We listen carefully to understand your needs and recommend actionable next steps.

Request a Meeting

Generative AI Research Guide

Created for a student audience by UW-Madison Science and Engineering Librarians, this Research Guide provides an overview of how these tools work and considerations when using them in course work, including tips for citing these tools.

Example Assignments and Lesson Outlines

The metaLAB at Harvard is curating assignments that integrate AI tools, contributed by instructors from around the world.

AI & Academic Misconduct

We recognize that the advancement of AI creates concern about academic integrity enforcement. While the focus of this guide is different, the following resources may be helpful for students and instructors.

Guidance for Writing

UW-Madison’s Writing Across the Curriculum staff have collaborated to draft guidance for instructors on supporting student writing practices in an age of AI, which will evolve as campus guidance and policy changes.

References & Further Reading

Acknowledgements

Thank you to Dorothea Salo, Dave Bloom and other UW Libraries staff, several anonymous commenters, and IDC team members for crowdsourcing and discussing these ideas.

This is an accordion element with a series of buttons that open and close related content panels.

How to Credit this Guide

This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. This means that you are welcome to adopt and adapt content, but we ask that you provide attribution to the L&S Instructional Design Collaborative and do not use the material for commercial purposes.

Example attribution: From Using Artificial Intelligence in the Classroom by the L&S Instructional Design Collaborative, licensed under the BY-NC 4.0 license.