AI Sandwich

Use AI tools for the beginning and end of an assignment, with the middle being grounded in human knowledge and expertise.

Assignment Details

Credit

AI Theme

Subject

Tools

Skills

Learning Objectives

Critically evaluate where in the research and writing process specific tools are and are not helpful

Instructions

Imagine you are teaching history, politics, international relations, or in any field where students could engage in first person interviews about a topic.

  1. AI as brainstorming partner
    • A student uses a large language model to brainstorm a list of research directions or interview questions about a given topic, dependent on the subject matter of the course. For example, in an international relations course, the student’s question might be: How has the proximity of Canada affected local economies in New England over the past fifty years? The student could ask the LLM: “Generate a list of 10 interview questions [for New England locals] that explore [how the proximity of Canada has affected local New England economies in the last fifty years].” The instructor should urge the students to push the model for more “out there” questions from the LLM, such as “how things might have been different if the border were completely closed” or “if God had swapped Mexico and Canada.”
  2. Human-to-human knowledge creation
    • In the second stage, the student should curate the questions they got with the LLM to the 5-10 they like the most and think will solicit the most interesting knowledge or feedback. They should then choose 2-3 people in communities relevant to their research question–could be a friend, relative, shop owner, etc.–anyone with a stake in, or knowledge about, the research question. The student can takes notes during the interviews, audio record them, or even use an AI note-taking tool to transcribe them (with permission, of course).
  3. AI as writing partner
    • The student uses an AI tool to convert interview notes into an outline. The student tweaks the outline, adds relevant framing, and asks the chatbot to draft an essay based on the outline. To encourage more authorial agency from the student, the instructor might ask students to generate three possible conclusions based on their notes, and the students can experiment with having the chatbot write in different styles or with different personae. The student may include one or all of these conclusions, but should assess how the truth value of each stacks up against what they learned from their own field research. Finally, the student can adjust the draft to match their opinions and submit the essay along with a summary or log of the prompt history used in the process. (ChatGPT makes it easy to export a log of your conversation as a zip file, though if you’re worried about feeding ChatGPT this human-gathered knowledge, you can turn off data collection in the settings.)
  4. Students should also include a final 1-2 paragraph reflection on the process that addresses some of the following questions:
    • What was helpful about using the AI tools in steps 1 and 3? 
    • How would the assignment have turned out differently if you (the student) had completed the full assignment without AI support? 
    • Did anything provided by the AI tool surprise you? 
    • Where did it underperform? Where did it overperform? 
    • Offer a theory about why it did well at certain tasks and poorly at others. If you were to share a piece of advice to fellow students about using AI tools in research assignments, what would your advice be? Please share your advice and include any caveats to it.

Related

AI Image Remixing

Students will work with an AI image mixer tool to explore the possibilities and limits of creative collaboration.

...

Debating the Ethics of Generative AI

Facilitate a debate between students about the potential harms and benefits of using generative AI tools in the classroom.

...