Recommended strategies for Generative AI in your class

Note: Instructors should think about Universal Design Learning when working on solutions to AI texts.

ChatGPT and AI Texts produce well written text gathered, but is not personalized. To reduce the use of AI text tools by students it is helpful to re-evaluate your assignments and ask:

Is the prompt generic? Generic prompts that ask students to simply define an item or list chronological events can easily be answered by ChatGPT and have little effect on critical thinking. Try expanding prompts that incorporate critical thinking or uniqueness in the response.
Is the assignment student centered? Students will want to write about something they are interested in. If the assignment has little connection to the student or it is not clear why the student should care about the assignment they may look for an easier solution. Try providing enough openness to the prompt so that students can focus on topics they care about or, in the assignment, explain the goals of the assignment.
What are the goals of the assignment? Review the goals you had when creating the assignment, could students (or yourself) view these goals as tedious or busy work? If so, change the assignment to have more robust goals or be clear on the usefulness of the goals. Finally, if in reviewing the goals you find the assignment does not actually promote them, change the assignment to better align with the goals.
Is writing the best way to reach the goal? Review if other ways for students to show learning could be used. AI text is hard to use if the final product is a video of the student or a in class presentation.

Students may come into your class being trained to think that writing is a formulaic practice that just regurgitates information. Discussing with the students on the utility of writing, how it can be used to not only convey information but formulate it, can lead students to reassess writing assignments.

Being open with the students about why AI text is not appropriate for this assignment not only shows respect for the students but also builds digital literacy.

AI text can incorporate false information – AI text generators use previously created material to form their answers. Their writing may be structurally correct, but the generators are not sophisticated enough to evaluate the information used to create the material. This leads to inaccurate or false information to be present.
AI text can be biased – AI tools have had a history of race and gender bias; AI text is no different. Depending on the material used to create the answer, the biases of the material can be amplified.
AI text can lead to homogenization – Expanded use of the AI text may be mechanically or grammatically correct but reduces the natural mistakes or creative writing found in human created text. This can reduce the growth of the language or incorporation of other cultures.
Academic honesty and AI – Talking with the students about academic honestly, why it is important, and why AI text may violate academic honesty provides students the opportunity to critically understand policies.