Ethical AI use for instructors

AI generated designs for your course or class meeting do not reflect your goals and aims as an expert in your field. It is unlikely that AI generated course designs will reflect your expectations for the kinds of work that you want students to do or the kinds of problems, activities, and sequences of learning that are crucial to developing the skills and knowledge students need to thrive in your course, in their program, in the university, or in life. A course is not just your intellectual property: it is an extension of your own thinking and your own identity as a scholar in a discipline into which you are inviting your students. You, a real, ethical, wonderfully unique human are inviting apprentices into your mental world. AI cannot do that.

What AI produces is always dubious and not the product of a thinking mind. AI functions by drawing on vectors of word frequency in the materials on which it has been trained. Some of those materials will likely be of poor quality and / or unrelated to your aims for students. Ultimately, a course design or class meeting generated by AI will be an average of sorts. The truly meaningful aspects of the work you want students to do will likely not be reflected in an AI generated design. The personal, inspirational design that is drawn from your expertise, passion, and humanity won’t appear in an AI generated plan for your students.

Put plainly, using AI has an environmental cost: excessive water and land use, consumption of electricity, and increased carbon emissions are some of the obvious costs of using AI to offload work that may be completely unnecessary. 

Unless you are certain that you are working with a secure platform, you might just be providing free content to an AI company. That content may be scrubbed and fed to another AI platform. Consider how you want your intellectual legacy to move beyond your control.

Finally, ethically speaking you must share with students that you used AI to craft some or all of your course design and class meeting design. Consider what kind of message this sends to students. If you want students to resist the pull of using AI to do their work in your course, it’s important to take the time to reflect on, design, and redesign the learning experiences you want them to have.

When we are teaching a course for the first time, we can’t always recognize the kinds of misconceptions students will bring in or what particular concepts or processes will be challenging for them. This kind of prompt can be helpful: 

I am teaching an undergraduate introduction to general chemistry course. Identify where students typically struggle in this kind of course and why. Please draw only on educational websites and research to formulate your response.

Keep in mind that you can ask this same question of your colleagues—in your department and in CATLOE! Alternately, take a look at the research on teaching in your discipline in higher education: there are entire journals and often books published about teaching in your field. The research will often pinpoint conceptual bottlenecks and disciplinary difficulties for you.

As you create your syllabus, your welcome letter, your assignments, your rubrics, or other important course documents, it can be very helpful to have another reader who can help you craft a document that you know will convey a positive tone and help your students understand your plans for their learning. This kind of prompt can be helpful: 

Please read my syllabus and make suggestions about how I could make the tone more inviting and warmer. I have a large percentage of first-generation college students and I want to make sure that the syllabus communicates that all students belong in my course and can succeed. Please use the guide to a warm and inviting syllabus that I have uploaded to make suggestions about my syllabus.

Keep in mind that you can ask this same question of your colleagues—in your department and in CATLOE!

Research suggests that well-structured classroom interactions that provide clear roles and clear “inroads” into interaction are a key for creating meaningful, inclusive learning (Hogan & Sathy). Initially picturing how this might work can be challenging. AI can provide you with some initial plans. A prompt like this can be helpful: 

I teach an undergraduate composition course and I want an activity that helps students give each other feedback on the first paragraph of a personal essay. Please create a peer review activity that is highly structured and requires students to focus on clarity and defense of a central idea with concrete examples of evidence.”

Developing a short case that is multifaceted and not easy to solve or resolve can be challenging. AI can provide you with some initial ideas. A prompt like this can be helpful:

I am teaching a course in counseling psychology, and I want students to discuss a short case study in class which requires them to apply the principles related to ethnic identity development, trauma, and family systems theory. Please write a two-paragraph case about a family from Ecuador who have recently immigrated to upstate New York. Give some details about the family and present a problem that my counseling psychology students would have to analyze using the principles I have identified.

When developing ideas for in-class activities, different types of “realia” have a variety of uses. A partially filled table of results, a set of DNA stains, a line of code, or a map of Victorian London might be what you need as the basis of an activity that requires students to make predictions, analyze an example, or apply a key concept. Developing these novel visualizations or concrete examples of a concept or process requires some thought and reflection. AI can provide some initial ideas and a sounding board as you decide what will help students do the particular thinking you want them to practice. A prompt like this can be helpful: 

I teach an undergraduate introduction to physics course and students are struggling to understand how magnetic force and electrical field interact. Find educational resources or webpages that explain the interaction between magnetic force and electrical fields with illustrations. These resources should use concrete, real-world examples that show the interaction of magnetic force and electrical field.

While we want students to wrestle with course readings and work hard to process those readings and develop a sound understanding of core concepts or processes, we often need to find ways to help student grasp those concepts. Figuring out why and how students struggle is a core aspect of teaching and AI can help you develop explanations and examples in response to student confusions. A prompt like this can be helpful: 

I teach an undergraduate introduction to anthropology course and students struggle to understand the Sapir-Whorf Hypothesis. My students come from three different language backgrounds: I have native English speakers (U.S. born), native Chinese speakers, and bilingual Spanish-English speakers. Please give me examples from the languages of U.S English, Mandarin Chinese, and Spanish (spoken in the Dominican Republic and Puerto Rico) that I can use to help them understand the Sapir-Whorf Hypothesis.

This can be very useful when you want to ensure you have more than one item that taps the skills you want to assess. This kind of prompt can be helpful: 

I am teaching Introduction to Statistics, and I want you to generate multiple choice questions related to the following topics: Types of Data and Data Collection, Frequency Distributions and Histograms, Graphs, Scatterplots, Correlation, and Regression. I would like these questions to challenge students and be written at the level of analysis and application. Please do not generate questions that merely require students to recall or regurgitate information. Generate three questions for each area.

Keep in mind that you can upload your textbook chapter or notes and specify that AI should generate novel problems related to that reference text. Also remember that you will need to review what AI produces carefully for errors or to see if the preparatory work you’ve had students do will actually prepare them for the items AI has generated. And remember to lean on human supports, too. Often a colleague can review test items and help you see where wording is needlessly complex or why students might be able to answer simply by guessing rather than doing the kind of higher-order work you might be aiming for. 

It isn’t always easy to develop a good case and AI can help you begin to develop your ideas. This kind of prompt can be helpful: 

I’m teaching a master’s level course in School Psychology and want you to develop a case that would require students to apply the frameworks of Bronfenbrenner’s Family Systems Theory, as well as the article on working with immigrant K-12 students. The case should involve a middle school student of color in a predominantly white school system. Provide a one-page narrative that involves a conflict between the parents of the student and a schoolteacher and administrator that the school psychologist needs to resolve. Please use the National Association of School Psychology web pages and peer reviewed research to create the case.

Keep in mind that you will likely need to rewrite what AI produces! You will need to add more nuance, more complexity, and red herrings so that students can really be challenged in the right ways with this case. You may find yourself completely rewriting what AI has produced, visiting the sites it draws its response from, or revising your prompt or asking for specific new details and ideas. 

Research suggests that the “transparency” of our assignment descriptions can highly impact the depth of engagement students have with the work we ask them to dive into (Winkelmes, Boye, & Tapp 2019). When we clearly and humanely communicate the “how” and the “why” of assignments, students are more able to do the pre-thinking and planning we want them to do, gather their inner, motivational resources, and slowly and thoughtfully work through the steps of longer projects. But it isn’t always easy for instructors to recognize all the work that is required for a project like a paper, a portfolio, or another complex performance. AI can help you review your description and suggest areas where you can make your assignment description more transparent and reveal the tacit work that is required to succeed. A prompt like this can be helpful: 

Please review my assignment description using the Transparent Assignment Design Template for Teachers available here: https://bdl2jezatgadvbda.public.blob.vercel-storage.com/pdf/Transparent%20Assignment%20Design%20Template%20for%20Teachers-P8byfuQgOHi8pcDuEskS71dvh5XbcR.pdf and the Checklist for Designing Transparent Assignments available here: https://bdl2jezatgadvbda.public.blob.vercel-storage.com/pdf/Checklist%20for%20Designing%20Transparent%20Assignments-rarHUjD2KFfDtncV5mie3uBuD7PMwp.pdf. Suggest areas where I need to increase the transparency of my assignment description.

Keep in mind that you can ask your colleagues, your students, a friend outside of academia, or the consultants at CATLOE to review your assignment description for transparency. It is humans—your students—who will be working to make sense of your assignment descriptions, so it always best to consider a review from AI only a first pass at an important document best analyzed by humans.