Integrating AI into Assignments

Getting Started with AI in the Classroom: Integrating AI into Assignments

As generative artificial intelligence (GenAI) continues to reshape higher education, you, as an IU instructor, are uniquely positioned to lead your students through a transformative experience. While we can all agree having generative AI thrust unknowingly upon us is a less than ideal situation, whether you’re now exploring AI’s potential to improve student learning outcomes, designing new assessments, or addressing its myriad ethical considerations, here are some best practices for integrating AI into your assessments.

Be Transparent

Transparency is a highly research-supported teaching and learning concept that demonstrates that explaining your reasons for why you do what you do in the classroom supports improved student learning outcomes. In the Age of AI, transparency has become even more important, as students navigate their differing expectations of their instructors. For example, one instructor might encourage the use of AI, and then a student might go to their next class where if they use AI they will be accused of academic misconduct. It is vital then that we as instructors be very clear with students about our course policy on the use of AI, even down to the assignment level. For example, in my first assignment I might ask students to write something about themselves—obviously I do not want this to be AI-generated, which would defeat the point of the assignment. But my next assignment asks them to critique AI outputs, an assignment that requires them to engage with AI. Therefore it is important that we are as transparent as possible with our students about how we would like them to use AI.

Give Options

Universal Design for Learning (UDL) is a systematic approach to preparing your course so that every student can participate fully. Rather than requiring students to ask for accommodations, UDL advocates for designing a course from the beginning to be accessible to every student. One of the key methods of UDL is offering students choices: one student might prefer to write a paper; another might prefer to do a presentation or a podcast. Ideally, offering students options gives them to their best opportunity to show what they have learned in your course in the best light. With regard to AI, asking students to incorporate generative AI into their workflows may prompt some students to feel excited while others might be more reticent. Also, individual students may find different places in their work where AI makes them more or less productive or effective. So, whenever possible, attempt to give students choices about how to use AI. For example, in the next section I describe how my first W131 assignment asks students to revise an AI-generated summary. But if they prefer to rewrite their final summary from scratch, I allow this as an option.

First Activity

As described in our generative AI in the Classroom modules (available for self-enrollment), your first assignment ideally explores generative AI in a way that helps your students understand that it is a tool that can support their learning, rather than a short-cut to avoid it. An easy way to make this clear is to find a prompt in your discipline that you can explore with your students to demonstrate both the advantages and limitations of using generative AI. For example, I wrote a prompt that produced a clearly insufficient explanation of cavitation. In my course (W131, Reading, Writing, & Inquiry) I use a prompt that produces a summary of a previously assigned essay. We talk about the strengths of the AI-generated summary, and then I proceed to describe all the ways it would fail to achieve a satisfactory grade in my course (for example they often lack specific detail and are missing descriptions of concepts or ideas that students previously identified as important in the essay). This will require you to engage with AI in your disciple—your area of expertise—to demonstrate for students the ways in which your course is valuable to them (even with ubiquitous access to AI tools).

Teaching Prompt Engineering

Once your students understand that simply submitting AI-generated answers will not lead them successful in your course, the real work begins: research is beginning to emerge that demonstrates that evaluating generative AI outputs helps students develop critical thinking skills, specifically metacognition, because not all prompts are created equal. While online articles will try to convince you this or that AI is better, what consistently produces better results is learning to write better prompts, a skill that students can (and probably will need) to learn. At this point I also consider this an equity issue: some students, often better resourced, have experience and some knowledge of how to write prompts. Other students may know little or almost nothing about how to write a successful prompt. Therefore students will need training or we will see achievement gaps. To be honest, it also often makes grading more difficult, as grading work submitted from badly written prompts is often more time consuming.

Therefore, as a metacognitive activity, consider teaching your students the basics of prompt engineering. This can easily be achieved by using a prompt formula, of which there are many. I use the example from our generative AI in the Classroom modules, but there are many others that may work just as well or even better in your disciple. But in my course I explain to students how to use the formula Role + Context + Task + Format. For example, in my course worksheet #1 is the AI-generated summary exercise described above. Then worksheet #2 guides students through using the Role + Context + Task + Format formula to write thesis statement examples, which I then ask them to critique. This gives them practice critiquing thesis statements while they also practice writing and evaluating prompts

 
Getting Assistance

To learn more about using generative AI in your classroom, consider self-enrolling in our generative AI in the Classroom modules, which are designed to help you get a foundation in generative AI and then get you started incorporating it into your classroom. You can also read my blog posts on AI, or watch our previous Faculty Showcase on AI assignments. If you would like a personal consultation on how you can incorporate generative AI into your classroom, contact the CITL with questions or for a personal teaching consultation on how to get started or to use generative AI to improve your learning outcomes.