Incorporating AI in Teaching: Practical Examples for Busy Instructors
Since January 2023, I’ve talked with hundreds of instructors at dozens of institutions about how they might incorporate AI into their teaching. Through these conversations, I’ve noticed a few common issues:
Faculty and staff are overwhelmed and burned out. Even those on the cutting edge often feel they’re behind the curve.
It’s hard to know where to begin.
It can be difficult to find practical examples of AI use that are applicable across a variety of disciplines.
To help address these challenges, I’ve been working on a list of AI-infused learning activities that encourage experimentation in (relatively) small, manageable ways.
Many of these examples were inspired by conversations with colleagues in the AI in Education Google Group I started in January 2023. I’ve found this community to be an invaluable resource as I try to address AI in my teaching and in the workshops I run for faculty. If you’d like to join the group, please mention the institution where you work and/or your reason for joining when requesting access here.
As I was working on this list, I noticed some common themes and tried to organize activities into discrete categories. Here are the eight categories I’ve identified so far (in no particular order):
Intentional Misuse
Text Revision
Critiques of AI-Generated Content
Creative Writing/Brainstorming
Role-Play
Presentation Preparation
Research and Inspiration
Instructional Design
Like so many things being written about AI in education right now, this document is a work in progress. Feel free to provide feedback in the comments or via email at daniel@danielstanford.com.
Intentional Misuse
We can think of intentional misuse as an “ethical hacking” approach where students misuse AI in order to expose shortcomings and threats. This type of activity can also foster discussion and critical thinking about the ethical dilemmas AI tools pose.
Example Activity: Use AI to create a misinformation campaign.
Example AI prompt: Create a 300 to 400-word article that explains why vaccines cause autism. Include links to at least two sources to support specific claims.
Example reflection questions for students:
Did the AI tool hesitate to answer your question? If so, how did you get around this?
What strategies were used to ensure the message would be believed and shared? (E.g., correlation as implied causation, appeals to emotion, credibility through association, repetition, confirmation bias, etc.)
Did the sources seem credible? Why or why not?
Part of my inspiration for this category came from a New York Times article on Finland’s approach to teaching students how to spot misinformation. While the article doesn’t focus on AI-generated misinformation, it mentions a middle-school teacher who, “…tasked students with editing their own videos and photos to see how easy it was to manipulate information.”
Text Revision
For activities where students revise or critique AI-generated text, consider how students will annotate their work. For example, your instructions might include one or more of the following guidelines:
Highlight text in yellow if at least 50% of the sentence was AI-generated without major revisions on your part.
Use track changes when revising AI-generated text so that the AI text is crossed out and your revised version is shown in a different color.
Identify at least one place where the AI-generated text is problematic. Highlight this text and use the comment feature to briefly explain why it’s problematic and how you would change the text to fix the issue.
Example Activity: Revise AI-generated text for a specific audience.
Example AI prompt: In 300 to 400 words, explain how a plane achieves the lift necessary for takeoff.
Student task: Revise the generated text and create two separate versions: one that can be easily understood by a 10-year-old and another that is appropriate for an adult learner.
Note: consider whether students should be encouraged or discouraged from using AI when revising AI-generated text. If AI use in revisions is discouraged, help students understand why this portion of the activity should be completed on their own.
Example Activity: Revise text for multiple formats/lengths.
Example AI prompt: Modify the following text for various audiences via a 400-word blog post, a single presentation slide, and/or a Tweet. [Paste text to be modified here.]
Example Activity: Summarize a text.
Example AI prompt: Summarize the most important points of the following article in five to 10 bullet points. Limit the summary to a maximum of 200 words total. [Paste article to be summarized here.]
Example Activity: Expand on an outline.
Example AI prompt: Take the following outline and expand on each point with one to two paragraphs of text. [Paste outline here.]
Side note: If you’re concerned about students using AI to generate a high-quality essay with minimal effort, I recommend reading, “Why I’m Not Scared of ChatGPT” by Christopher Grobe. The article provides a thoughtful take on the value of human intervention in the writing process. It also shows the amount of skill that can be required when revising AI-generated text if we’re thoughtful about the design of writing assignment prompts.
Example Activity: Revise AI-Generated text to use specific rhetorical/literary devices.
In, “Alternative Strategies for Artificial Intelligence in the Writing Classroom,” Abriana Jetté provides a fantastic list of ways students can modify and add on to AI-generated texts. Examples include:
Encourage students to play with litany or anaphora, perhaps using repetition as a replacement for “however,” “therefore,” or “additionally.”
Tell students to replace any sentence with a question.
Incorporate the craft of blackout poetry directly onto the AI generated work, a process which helps students with concision as they narrow down the words with each thick, black line.
Ask students to write a metaphor or simile about the AI-generated essay.
Critiques of AI-Generated Content
Example Activity: Critique AI-generated arguments.
Example AI prompt: Should governments provide tax credits for electric vehicles?
Student task: Paste AI response into a document and use “track changes” mode to critique/revise.
Side note: In March 2023, I hosted a workshop at the SUNY Online Summit where I shared an example of how ChatGPT responded to the electric-vehicle-tax-credit prompt above. You can view ChatGPT’s response to this and other workshop attendees’ prompts at bit.ly/chatgpt-examples.
Example Activity: Analyze examples of bias in AI output.
Example prompt: Write a scene in a movie script where people in specific professions interact (e.g., a doctor/nurse, pilot and flight attendant).
Student reflection question: What gender did AI assign to each role? How did this reinforce or contradict common stereotypes?
Additional example assignment: Analyze how AI handles racial or gender representation depending on prompt wording. See examples at bit.ly/bias-in-ai-images.
Example Activity: Critique or defend AI self-censorship.
Example prompt: Provide feminist interpretations of verses X and Y from religious text Z. (Inspiration for this example came from Maha Bali, who noted in early 2023 that ChatGPT refused to answer questions related to feminist interpretation of Qur’an passages.)
Student task: Identify a question that AI refuses to answer and justify why the tool’s self-censorship is positive or negative in this case.
Creative Writing/Brainstorming
Example Activity: Write a pitch for a TV show.
Example prompt: Write a pitch for a TV show in the style of Undercover Boss, but with University Presidents going undercover at their own schools. See example output at bit.ly/undercover-boss-pitch.
Example Activity: Create or refine a script.
Example prompt: Write a scene from a movie where the following characters interact. [Insert character descriptions here.]
Example Activity: Help with product development/marketing.
Example prompt: Generate five names and taglines for a product that does X.
For a similar example, see ChatGPT’s attempt to come up with a clever name for a store that sells exotic salts at bit.ly/ai-biz-names.
Role-Play
Example Activity: Ask AI to play the role of X.
Example prompt: Play the role of a flood warden with 20 years of professional experience and a personal experience with flooding.
Student task: Interview ChatGPT as it plays this role. Analyze the output to refine interview questions before meeting and interviewing a real flood warden.
Source: Katherine Walsh, University of Chester. See details at “100+ Creative Ideas to Use AI in Education,” slide 97.
Example Activity: Develop role-play scenarios for use in class.
Example prompt: Write a role-play scenario in which a manager meets with an employee to discuss a specific micro aggression she made in a conversation with a coworker. See example ChatGPT output at bit.ly/ai-role-play-examples.
Presentation Preparation
Example Activity: Create a presentation outline.
Example prompt: Create an outline for a presentation on the most common mistakes healthcare workers make that jeopardize patient safety. See example ChatGPT output at bit.ly/ai-patient-safety.
Example Activity: Create presentation slides from an outline.
For a list of AI-infused presentation tools, see geekflare.com/best-ai-presentation-makers.
Research and Inspiration
Example Activity: Assemble a list of X that meets specific criteria.
Example Prompt: List five websites that have good interface design. Limit the list to mid-sized companies that are not in the Fortune 1000.
See my LinkedIn post for an example of how Forefront.ai responded to a similar request and my critique of the output.
Example Activity: Create a “best-of” list and justify its methodology.
Example Prompt: Create a list of the 10 most artistically significant independent films of all time. Justify why each deserves to be on the list. Include three films that are also strong contenders and explain why they were not included in the top 10.
Instructional Design
The prompts in this section could be used by you (the instructor) to generate instructional materials and assessments. Alternatively, they could be used by students as part of an assignment where they develop similar materials to demonstrate their mastery of the subject matter.
Example Activity: Generate guides, checklists, plans, or instructional materials.
Example prompts:
Create a nutrition plan for a diabetic patient that incorporates the following foods. [Insert foods here.] Provide recipes for breakfast, lunch and dinner for each day of the week.
Write 20 quiz questions that test past-tense verb conjugation for ESL learners with moderate fluency.
Create a checklist of dos and don’ts for an entrepreneur documenting business expenses for tax purposes.
Generate five vegan recipes that use ingredients X, Y and Z.
Other Suggestions?
What categories, examples, and resources would you add to this list? Are there any words of caution or encouragement you’d like to share? Feel free to weigh in with a comment or email me at daniel@danielstanford.com with your thoughts.
Great article, thank you!
Thank you for your work in this space. Lots of good ideas here.