Designing assignments
Designing assignments effectively relies on the same principles as before the rise of natural language processing tools, with a few added considerations explained in this section.
As you begin to (re)design assignments in your course, we encourage asking yourself the following six questions (adapted from Derek Bruff, former director of the Vanderbilt University Center for Teaching, as written in his blog):
- Why does this assignment make sense for this course?
- What are specific learning objectives for this assignment?
- How might students use AI tools while working on this assignment?
- How might AI undercut the goals of this assignment? How could you mitigate this?
- How might AI enhance the assignment? Where would students need help figuring that out?
- Focus on the process. How could you make the assignment more meaningful for students or support them more in the work?
Overall strategies
Below are some practical tips and strategies that prioritize authentic, original writing no matter how foregrounded you want AI to be in your course.
Connect students with the class context and each other
- Design writing prompts that make reference specific to your class material. For example: “Use at least two theorists discussed in class to support your answer.”
- Ask students to draw on their responses to each other in the classroom or in classroom contexts, like discussion boards or blogs.
- For example: In David Lipscomb’s “Writing and Culture” class, each student maintains a blog (via CNDLS Course Sites), posting and responding to peers’ posts every week.
- Assign personalized writing. This approach may help you get to know students’ writing in more detail so you can recognize their style and tone.
- Tailor your writing assignments to focus on a specific rhetorical context, as AI struggles to focus on the audience in tandem with tone. For example, review this introductory assignment in a Georgetown computer science course.
Facilitate reflective learning
- Incorporate real-time activities and assessments in your course. Read more on our Active Learning page about how to do this.
- Provide incentives for the learning process as well as the product(s). If a perfect product (test, paper) is the only way to receive an A, students are more likely to consider cheating.
- Consider assigning students a process statement (a paragraph that describes the process they used to develop their final product) to submit along with their assignments. How did they use generative AI, and what did they learn from it?
- Review your grading criteria and rubrics to make sure you’re setting your students up to adopt strong learning strategies.
Work directly with AI
- Have your students use ChatGPT to answer a prompt and ask them to respond to the answer provided. Where does the auto-generated information succeed? Where does it fail? Where does it not understand the nuance or depth of the question?
- Incorporate AI into a drafting process:
- Have AI write a first draft and then ask students to edit it or vice versa
- Have students write the drafts and ask AI software to edit it. A key question and skill may become teaching students to coach AI to generate quality writing.
- Teach students to ask good questions of AI tools. Understanding how generative AI tools work could benefit students in the long run, as these technologies continue to develop. Asking good questions is at the heart of scholarship—how might generative AI begin to augment research processes?
Assignment examples
Below, see examples of assignments Georgetown faculty have adapted and newly created to account for generative AI’s capabilities.
Adapted assignments
| Faculty | Original assignment | Adjusted assignment |
|---|---|---|
| Sherry Kao (Philosophy) | “Present and discuss a philosophical debate, concepts, or theories with your friends or family in the format of a podcast interview, email exchanges, a TED-Talk style video, or social media.” | “Use ChatGPT to generate an essay. Identify strengths and weaknesses of the essay, identify any knowledge gaps and fill the gaps. Write a meta-cognitive reflection on whether and how using ChatGPT could enhance their learning.” |
| Faculty | Original assignment | Adjusted assignment |
|---|---|---|
| Nicholas Lovegrove (McDonough School of Business) | “Write a paper about the respective business models of two businesses that compete.” | “I’ve asked ChatGPT this question and I want you now to critique what is said. Identify where you think it’s wrong. Identify where you think it’s right and improve upon what it’s done.” |
New assignments
| Faculty | Assignment |
|---|---|
| Rebecca Helm (Earth Commons) | Prompt ChatGPT to write a four-paragraph essay, at the college level, on an expert-level topic of your choosing related to this class. Write a companion piece analyzing each paragraph of ChatGPT’s essay. For each paragraph, you will comment on what it got right (and the source of that information), what it got wrong (with sources for the correct information), and where it may be missing important information. Submit your prompt, the ChatGPT essay, and your companion piece, as a single document. |
| Janet Gomez (Liberal Studies) | Use AI tools such as hotpot.ai or DALL-E to generate images of female rulers. Input the characteristics you believe female rulers should have. Goal: Students will examine their perceptions of a female ruler, share their perceptions using visuals, and reflect on their biases on female leadership. Key Questions: Can AI be biased in how it produces images of human figures? How can we take advantage of the convenience of AI tools without reinforcing biases in students’ learning? |
| Nathan Schneider (Computer Science) | Schneider asks students to directly interact with ChatGPT, and then asks: “Do you think it can be useful to your learning (not for cheating but for improving your understanding of concepts in a course)? Why or why not?” Read the full assignment. |
| Nathan Hensley (English) | Lecture: Once Prof. Hensley revealed his lecture was AI-generated, he distributed the text of the lecture and asked students to consider 1) the issues with what an NLP suggested for a survey course and 2) how that raises questions about what ‘should be’ covered in an intro-level class. He then asks students to consider the following sources: 1) Matthew Kirshenbaum, AI Is Ushering in A Textpocalypse, 2) Ian Bogost, ChatGPT is Dumber Than You Think, and 3) Paul Fyfe, How To Cheat On Your Final Paper (video). After the lecture and reading, he assigns the following: “Study ChatGPT, test it out, and read up about large language models and AI writing technology. Write some stuff with it. Answer the question: Is it writing?” |