Appendix C: Incorporating AI into the Classroom

Below, we offer five opportunities for blending AI into your writing classroom.

Option 1: What Is “Writing” in the Age of AI?

Learning objectives:

• Identify the major functions of ChatGPT
• Reflect on the functionality of ChatGPT as a writing technology, reinforcing specific beliefs and values about writing
• Identify specific use cases for ChatGPT
• Identify at least one limitation of ChatGPT in light of the definitions of writing as a rhetorical process

Scope/timeline: one or two class sessions

Prep work required/recommended: 

• In a previous class session, read/discuss the Naming What We Know chapter by Brooke and Grabill, “Writing is a Technology through which Writers Create and Recreate Meaning.”
• Review the concept of writing as rhetorical: as goal-oriented communications between or among interlocutors in response to specific exigencies that are socially structured and recurring
• Set up a ChatGPT account

Activity: I Do/We Do/You Do

Critical Question: How is the concept of “writing” changed by generative AI tools?

1. As a class, identify specific use cases for ChatGPT to help with a writing task. Have students consider the rhetorical situation and the attendant genre that would be produced, and list the examples on the board, working together to revise the scenarios to be more specific about the rhetorical context: speaker, audience, message, purpose, etc. Examples might include generating a personal or business email to communicate a particular message, composing a specific class assignment for a particular discipline, writing a cooking blog, writing a response to a particular text, etc.
2. In small groups, students will choose one example and develop a prompt to have ChatGPT complete their selected task. The prompt should identify the genre (email, research paper, blog, etc), context, audience, speaker, and purpose of the writing.
3. Students will analyze the output for its rhetorical effectiveness and be prepared to share their findings with the class. Consider ethos: Does the output have a sense of voice that seems fitting and appropriate in the given social context? How does it establish authority or credibility? Also consider histories: How does it incorporate or respond to specific social contexts and histories, such as class content that would have been covered, communication norms and expectations, etc. Lastly, consider content: Does the AI have the information it needs to be accurate? Does it have the genre knowledge to give output in the expected form?
4. Discuss the findings as a class, seeking to identify patterns in the kinds of tasks the GPT appears to be able to do effectively, and what kind of interventions are needed to improve it or adjust its output
5. In small groups again, craft a prompt for revision to transform the output: to take on a different genre, address a different audience, be composed by a different writer, in a different context, or for a different purpose. Prompt the GPT to revise its previous output according to the new parameters.
6. Individually, reflect on the original output, the revised output, and how they compare to what you might have composed yourself. What are the affordances and limitations? Is what ChatGPT or other generative models produce “writing”? Is it “rhetorical”?

Option 2: Co-Creating a Course Policy for AI Use

Learning objectives: 

• Analyze the generative AI output created in response to class assignment prompts
• Propose guidelines for a course policy around the use of Generative AI

Scope/timeline: half of one class session, with follow-up in the proceeding class session, ideally at the beginning of the term

Prep work required/recommended for instructors: 

• Feed prompt(s) for early class assignment(s) to the AI tool, and then copy the output onto the slide deck

Prep work required/recommended for students: 

• None required

Discussion:

• Opening:
  • Ask students to review AI-generated response(s) to class assignment prompt(s), then ask how reviewing the response makes them feel
  • Small Group Discussion:
  • In groups of 2-3, have students discuss the following questions and add their responses to a collaborative document similar to the example below.
    • In what scenarios do you think it would be useful to draw on Generative AI tools for the work of our class?
    • In what scenarios do you think it would NOT be useful to draw on Generative AI tools for the work of our class?
    • In what scenarios would the use of Generative AI in our class constitute a breach of academic integrity?
    • What are the things you want me, your instructor, to do when it comes to the use of Generative AI tools?
    • What are the things you want your fellow students in this class to do when it comes to the use of Generative AI tools?
    • What are your other instructors saying about Generative AI tools?

Follow-up:

• Review students’ responses to the questions on and draft a course policy that reflects their ideas.
• Present the draft in the next class session and solicit feedback.
• Revise based on feedback and add policy to syllabus.

Option 3: Generative AI & Summarizing Scholarly Sources

Learning objectives:

• Compare human-written summaries of scholarly sources and generative AI-created summaries of scholarly sources
• Identify and apply strategies to improve AI’s summaries of scholarly sources
• Evaluate generative AI’s utility in the process of summarizing scholarly sources

Scope/timeline: one or two class sessions, ideally in preparation for writing a literature review

Prep work required/recommended:

• In the previous class session, we read a scholarly article as a class and, together, wrote a summary of it using best practices for summarizing complex texts
• Before this class session, select several scholarly articles and, in small groups, have students read one of the articles
• Set up a ChatGPT account

Activity: I Do/We Do/You Do 

• Opening (I Do):
  • Compare the summary of the scholarly article you wrote as a class during the last class session to a ChatGPT-generated summary of the same article
  • Reflect on the prompt used to elicit the ChatGPT-generated summary and identify ways to improve the prompt
• Activity (We Do):
  • In small groups with other students who read the same scholarly article for homework, draft a summary of the article. After that, have ChatGPT generate a summary of the same article. (Can use multiple prompts to generate ChatGPT summary.) Reflect on the differences between the human-generated version and the ChatGPT-generated version, using the following templates to take a critical perspective on the ChatGPT-generated summary (borrowed from Anna Mills, “Template Phrases for Critiquing AI Outputs [Website]”):
    • This sounds plausible because it ________, but it doesn’t really make sense because ________.
    • This sounds good, but it doesn’t really fit the purpose. What we are looking for is __________.
    • This is inaccurate because __________.
    • The AI seems to have misinterpreted __________.
• Class Reflection/Discussion:
  • Based on your experience with this activity, what do you see as the strengths or limitations of human-written summaries and/or of AI-generated summaries?
  • Is it worth it to use the ChatGPT-generated summary as a starting point, or would you prefer to start from your own summary? Why?
  • What are some ways that you think you could use ChatGPT in the summary process? What are some sample prompts or types of prompts that might help you with your literature review?
• Optional Follow-Up Low Stakes Assignment (You Do):
  • Write a summary of one of the articles that you are planning to incorporate into your literature review. Have ChatGPT write a summary of the article (include the prompt(s) you used to elicit its summary as well as the summary itself). Compare the two and reflect on ChatGPT’s utility (or lack thereof!) in the process of summarizing scholarly sources.

Option 4: Formulating Research Topics and Questions

Learning Objectives:

• Identify a research topic, moving from a broad area of interest to a more focused, specific area of inquiry.
• Learn to use ChatGPT as a tool for brainstorming and narrowing down research topics.
• Develop a deeper understanding of what prompt engineering is and how it affects responses generated (creating effective prompts, role-playing).
• Evaluate the benefits and limitations of using ChatGPT to identify and narrow down research topics.

Background:

This activity was originally designed to follow a unit in which students analyzed materials from the archives of The Santa Clara student newspaper and the Multicultural Center. In this context, students analyzed former students’ perspectives on topics of interest (such as immigration, mental health, sustainability, etc.), shedding light on how the university community has historically engaged with these issues. The goal was to generate ideas for research topics that hold significance for past and present members of the university community. However, the activities in this module can also be adapted for a range of assignment types and sequences that help students identify a topic of interest.

Resources:

• BIPOC SCU Student Archival Resources [Website] (SCU Archives & Special Collections)
• Introduction to ChatGPT & Generative AI [Website] (Tidewater Community College)
• Strengths and Limitations of Using AI for Students [Website] (Tidewater Community College)
• The Importance of Prompt Engineering [Website] (Tidewater Community College)
• Best Practices for Prompt Engineering [Website] (OpenAI)
• Innovation through Prompting [Website] (Ethan Mollick)

Preparation:

• Prior to this activity, students should have engaged in analytical work that helps them to identify a broad topic of interest, ideally one that holds significance for members of the university community and/or a particular field of study (here is a resource [Website] that can be used for this purpose).
• Instructors can have students review the #ImmigrationSyllabus [Website], considering the keywords featured each week and observing how a topic, such as immigration, can be broken down and explored through multiple lenses.
• Students should have a ChatGPT account.
• The class should have an existing policy around generative AI use in place.

1. Brainstorming keywords and subject areas

Working individually or in small groups, students will

• Generate a list of keywords and subject areas (categories, themes, fields of study, etc.) related to their selected topic.
• Use a concept map to organize, break down, and make connections between keywords and subject areas, as well as generate further ideas.

2. Using and evaluating ChatGPT as a tool for brainstorming

• Ask students to prompt ChatGPT to generate a list of key terms and subject areas relevant to their topic.
• Have students compare the ChatGPT-generated lists with their own lists and evaluate the results. What similarities and differences are there? Are there any terms or subject areas from the ChatGPT list that helped you expand your own thinking about your topic? Were there any terms that seemed problematic (irrelevant, biased, etc.)?
  • Example of irrelevance/bias: When prompting ChatGPT to generate a list of key terms for the topic of immigration, the results included the phrase “brain drain.”
• Discuss the concept of prompt engineering [Website]. and six strategies [Website] recommended by OpenAI.
• Have students apply one of the strategies, such as role-playing, to refine or re-create their prompts (e.g., “Acting as a historian, identify the most important issues related to immigration in the US over the past 25 years”).

3. Reflect

Have students reflect on their experiences with using ChatGPT and the strengths and limitations of using it to develop their research topics.

• What advantages did you find in using ChatGPT as a tool for brainstorming and narrowing down research topics?
• What limitations or challenges did you encounter while using ChatGPT in this context?
• Based on your experience, what additional resources or strategies would you use to identify and narrow down research topics?

4. Moving from Topics to Research Questions

Students will draw on the terms and topics they identified to help them formulate and refine a research question, using the following resource and activity:

• Resource: Developing a Research Question [Website]
• Importance of Prompt Engineering
  • Think critically about your prompt, as what you input determines the output. If you simply ask ChatGPT to revise your question, the likely result will be a rephrased version of the original question, which may not be very helpful.
  • Instead, start by asking ChatGPT to list underlying factors, keywords, or subtopics that are relevant to your question. Next, ask it to revise the original question to incorporate one or more of these elements.
  • Remember to put the revised question in your own words—so that it reflects your own unique voice. Don’t let ChatGPT do the writing for you (the goal is to use this as a tool for revision, as part of a socially just and culturally sustaining language and writing practice).
• Activity:
  • Goal: Revise your research question for
    • clarity
    • scope (using who, what, when, where, why questions),
    • and focus (identify relationships, causality, or underlying factors).
    1. Ask ChatGPT to suggest revisions for your question based on these goals.
    2. Use at least 3 different prompts (include a different goal in each prompt).
    3. Revise your question based on these results (don’t just copy word for word).
    4. Share your original question and revised question. Briefly describe how you used this exercise to revise your question.

Option 5: Revising an Essay with Generative AI

Learning Objectives:

• Identify multiple ways that generative AI can be used as a tool in the revision process and assess its effectiveness as a tool.
• Identify multiple strengths and weaknesses in generative AI’s ability to compose texts.
• Identify features of their own writing style/personal voice and gain confidence in their abilities as writers and researchers.
• Develop and implement a revision plan for their essays.

Scope/timeline: One or two class sessions

Prep work

• Students should have a draft of an essay for review. Preferably one for a topic that they have done some research on/have some expertise in.
• Students should have a generative AI account (ChatGPT, Claude, CoPilot, Gemini, etc.).
• The class should have an existing policy around generative AI use in place.

Activity:

• Begin by asking students how they usually review essay drafts (read it over, ask someone else to read and comment, run through a checker like Grammarly or AI, etc.).

1. 1. What do they think is effective/ineffective about those approaches?
   2. What do you like/dislike about your own initial essay drafts? About revising your essays?

• Ask students to prompt the AI to generate a short essay (250-500 words) on their topic.

1. 1. What strengths does the output have based on your understanding of the topic/based on the knowledge you have gained researching your topic?
   2. What are the weaknesses?
   3. How would you describe the AI’s writing voice/style? What is good/bad about the way AI writes?

• Have students upload a draft to an AI (or copy/paste it in the prompt box) and ask it to revise for grammar and style.

1. 1. Have them compare the versions (In Word, under Review > Compare; In Google Docs, under Tools > Compare Documents)
   2. Pair up and review the comparisons together. What do they notice about the comparisons? Are there any patterns in what the AI changed from their original document? Are there any changes they disagree with? How does it change their personal writing voice from the draft? Is that a positive or negative change? Why?

• Have students continue the AI conversation by submitting the prompt “Can you suggest any areas where the content needs revision?”

1. 1. Pair up to review the response. What aspects of the AI’s feedback surprise you? What doesn’t surprise you? Which suggestions do you plan to take or reject, and why?
   2. Create a list of concrete steps you can take to address the AI’s suggestions.
   3. Consider letting students pursue the conversation further, especially if the AI prompts them to ask for specific suggestions or elaboration on its content recommendations. Ask them to review and reflect on the responses again.

• Have the students prompt the AI with the questions listed below. Each time, they should review the responses with their partner, note what stands out/surprises them, note what confirms their own perceptions of their draft, reflect on what they will accept/reject from the suggestions, and develop a list of specific steps to address the suggestions.

1. 1. “What are the strengths of my essay?”
   2. “How might I revise my essay further to preserve my personal voice?”
   3. “How should I prioritize my revisions?”

• Have them assemble the notes they took after each prompt and develop a plan for revising their essay based on the AI’s output.

Reflect:

1. How effective/ineffective is AI as a revision tool?
2. What did chatting with AI reveal to you about your own writing?
3. How do we need to acknowledge the use of AI as a revision tool to be in line with our class policies? (Or, do class policies need revision?)