1 of 59

AI in Your Discipline: Gathering ideas, resources, and approaches tailored to your subject area.

Anna Mills

AI FAST Workshop 2

College of the Siskiyous

October 24, 2025

Licensed CC BY NC 4.0

​

2 of 59

Welcome! What to expect

  • Choosing an experiment in AI pedagogical integration
  • Looking for ideas in your discipline
  • Logistics: Choosing tools and managing AI assignments
  • Giving students context on AI and encouraging skepticism

3 of 59

Housekeeping

  • Slides: Use the QR code or link: https://link.annarmills.com/FAST2
  • Feel free to put questions or comments in the chat or raise your hand in person.

4 of 59

A note on preventing AI misuse

This always comes up whether it’s the topic of an AI workshop or not!

5 of 59

AI in pedagogy: How do we decide where to start?

6 of 59

My discipline-specific goals when I consider AI integrations (modified from the principles in my last presentation)

  1. Help students meet existing learning outcomes
  2. Address a need or challenge I’ve encountered that couldn’t be addressed well before.

7 of 59

Example 1 of meeting a need:

Grammar and Style Explainer Bot

I mark 3 problematic sentences on each essay and ask students to correct and explain them using OER references. But students often need more tailored explanation of how particular grammar and style concepts apply to their sentences.

This isn’t a need where relationship is key; time with human tutors is better spent on ideas.

So I wrote a prompt to produce tailored explanations without correcting the sentence for the student.

8 of 59

AI examples, as mentioned in the last presentation

The need: fresh examples of course concepts tailored to student interests. (It’s prohibitively time-consuming to keep writing these.)

​

With AI, generate live and critique together or choose generated examples ahead of time or assign students to prompt for examples and critique or iterate until they are satisfied.

9 of 59

Assumptions Analyzer Chatbot gives examples related to meme chosen by the student

​

Need: examples of assumptions related to an argument each student is interested in and knows well. Students have analyzed an argument in a previous essay, so I want them to see how assumption analysis can help them go deeper with that particular argument. Often they miss many of the assumptions.

​

See the assignment on Canvas Commons (log in first so the link works).

10 of 59

Generate and critique examples of class concepts and how they apply to students’ real-world concerns. 

​

  • Prompt: “Give an example of the use of regression analysis on data relating to anxiety and screentime.” (sample without sources/sample with sources)
  • Prompt: “Give an example related to diet where correlation data is misleading and does not really reflect a causal relationship.”
  • Prompt: “Give an example of a patient whose experiences point to the limitations of current DSM diagnostic categories.”

11 of 59

12 of 59

In the AI Validation activity, students evaluate competing accounts of how to think about a problem

“Take a challenging problem on the topic your class is working on. It should be a problem where there is either one right answer or where there are answers that are “more right” than others.

Give the problem (and any context you would usually give to students) to an AI tool. Ask it to generate multiple solutions and explain its reasoning.” - Robert Talbert, AI Validation, AI Pedagogy Project

13 of 59

AI feedback, as mentioned in the last presentation

Need: Immediate on-demand feedback. I know how important it was to me to get feedback on the spot from my parents the night I was writing the essay (in addition to tutors and teachers and peers at other times); not everyone has family members ready to do that.

14 of 59

Simulations

  • What decision-making processes and human interactions do students need to practice more in your field?
  • Would it help to practice these with a chatbot before practicing live with humans?
  • Consider casting the chatbot in a role and asking students to practice particular skills
  • See the sample interview practice assignment from Esperanza Román Mendoza

15 of 59

16 of 59

Ideas in your discipline

​

17 of 59

CUNY Discipinary Toolkits for Education, Health, Humanities, Social Sciences, and STEM

18 of 59

Filter the AI Pedagogy Project assignments by subject

19 of 59

For the CTE folks: ChatGPT DeepResearch reports on your pedagogical use of AI in your fields (see the chat transcript)

​

20 of 59

Conference programs and disciplinary listservs as sources

Where do people talk about pedagogy in your discipline?

  • The last two years of conference programs will likely have a lot of approaches to integrating AI that have at least been through the conference review process…
  • Listservs or private Facebook groups
  • LinkedIn or Bluesky (e.g. search on Fire science AI pedagogy)

21 of 59

Nurturing student confidence and skepticism when it comes to AI

​

22 of 59

Possible AI-related goals of AI integrations (modified from the principles in my last presentation)

  • Help students build skepticism of highly plausible AI-generated outputs
  • Help students build confidence and metacognitive awareness

23 of 59

In a video or in class, model skepticism and self-awareness of process as we review AI outputs

  • skepticism and verification: “Hmm, does the source really say that? Let’s see…”
  • error spotting: “Hey, that doesn’t make sense.”
  • bias spotting: “Notice it’s slanting this toward AI hype. I didn’t ask for that.”

​

24 of 59

One approach to cultivating skepticism: ask for contradictory AI responses

Invite students to evaluate competing contradictory AI answers to a course-related question. (See the Contradictory Chatbot prompt)

​

25 of 59

For me, contradictory feedback on my own writing has spurred more critical reflection and sense of agency at times

​

26 of 59

When I asked for feedback, Contradictory Chatbot’s three suggestions for revision were indeed in tension in ways that helped me clarify my priorities

​

​

  • “Focus on streamlining the length and structure. The paper is too long and meandering…”
  • “Strengthen the theoretical framework first. The paper needs a much more rigorous analysis of how metaphors shape understanding of technology,”
  • “The top priority should be adding more concrete examples of AI outputs that illustrate the difference between hallucinations and mirages. The paper is too abstract and conceptual”

The right track…

27 of 59

Scaffold prompting strategies

In a chat assignment, you may want to offer an introduction to prompting strategies. Here’s mine, and in it you’ll find links to several more:�Getting the most out of AI (prompting): Describe what you want and keep trying from AI and College Writing: An Orientation.

Photo by Miguel Á. Padriñán on Pexels.com:

28 of 59

Main messages about prompting

  • Be specific (for example, tell it context, role, tone, style, examples)
  • Ask it to ask you questions about what you want
  • Get AI to take it step by step
  • Push the AI to do better
  • Use what you know (most of what you learn in school helps you prompt)
  • Be playful and curious

29 of 59

PAIRR includes guidance for the follow-up chat

“For completion credit, please continue the conversation by chatting back at least twice. Ask absolutely anything about your writing. Ask for clarification or explanation, push back if you disagree with feedback, share a particular challenge, whatever. See Getting the Most Out of AI Feedback.”

​

30 of 59

In MyEssayFeedback, students see this guidance below the AI feedback

31 of 59

Dropdown menus offer strategies and phrases to get more out of AI feedback

32 of 59

Give credit for reflection as part of the assignment

If we give homework credit for reflection on skeptical student reflections on AI outputs and how these were useful or misleading, we’re sending the message that some of the learning happens when they step back from AI and center their own judgment.

I find these reflections really interesting to read and respond to.

33 of 59

The PAIRR Process: A Metacognitive Model

​

34 of 59

A few PAIRR reflection questions

  1. What insights or perspectives did you get from AI? How do you feel about the feedback you've received on your writing? What advice do you want to take and why? What advice do you disagree with and why?
  2. What, if anything, was helpful about asking AI follow-up questions?
  3. Given the feedback you received, what steps are you planning to take as you revise?

From the The Peer & AI Review + Reflection (PAIRR) Packet

35 of 59

What are you thinking you will try?

36 of 59

Take 5 minutes to explore

And then share in the Mentimeter: https://www.menti.com/alz5oqd6jtva (in the chat)

  • How are you thinking of inviting students to engage with AI in your course?
  • What are you excited about?
  • What challenges might arise?

37 of 59

38 of 59

39 of 59

40 of 59

Logistics of assigning AI interaction

​

41 of 59

Choosing your tool

FERPA compliant:

  • Copilot through College of the Siskiyous Microsoft accounts (student or faculty)
  • PlayLab custom bots (more about them later)
  • Some discipline-specific AI ed tech apps like MyEssayFeedback (I’m a volunteer advisor)

​

Another approach: Offer a FERPA-compliant option but let students choose which chatbot they will use. Often prompts will work in many platforms.

42 of 59

For feedback, consider the Peer & AI Review + Reflection packet of assignment materials

​

43 of 59

How students can share chat sessions

Ask students to use the share function in the chatbot app and copy the link into their assigned reflection

  • ChatGPT, Claude, Copilot or Gemini: use the share function
  • PlayLab chatbots: “share this conversation” creates a URL with the chat transcript. I have found this only works when the user is not chatting while logged into a PlayLab account (hoping they’ll fix that soon).

OR copy and paste into a word processing app and share that document.

44 of 59

How students can comment on chat sessions

45 of 59

Screenshot from an annotated PlayLab.ai chat in Google Docs

46 of 59

Working with sources

Do you want AI to reference course materials?

If you have permission or the license allows for it (as in OER licenses), you can upload a file or ask students to upload a file when they chat.

I uploaded particular OER chapters on grammar and style to the Grammar and Style Explainer Bot and also put links to them in the prompt so the bot could provide students with the appropriate link (Playlab bots don’t browse).

Tip: in some cases you might want to ask AI to quote from the materials.

47 of 59

Create your own custom bot?

  • A custom bot is really a long prompt you put behind the scenes so students don’t have to copy and paste it. (Sounds fancy, though!)�
  • PlayLab is a nonprofit whose purpose is to allow educators to make custom bots. There may be a waitlist for the training, so if you aren’t able to get into a training, you can start out in the workspace I created for educators. They offer online guides.

48 of 59

How to offer alternatives to AI assignments

  • Students may prefer not to use AI because of ethical or privacy concerns.
  • Offer an alternative assignment, such as a reflection on an additional tutoring visit or a short reflection on feedback from friends or family.
  • I ask students doing the alternative assignment to submit text responses to reflection questions.
  • I describe the alternative in the same assignment instructions for easier grading where possible. Where it isn’t, I create a zero point alternative assignment where they can earn bonus points that make up for missed points on the AI assignment.

49 of 59

Sample open-ended AI research reflection assignment with alternatives

  • My discussion assignment: Learn about AI research assistance, experiment if you choose, and reflect on Canvas Commons
  • I assign this after I assign library workshops on using the databases, traditional research explorations and reflection on sources, and some drafting
  • Even here, I emphasize ways they could get more out of traditional search, such as working with a librarian
  • Most students used Stanford’s STORM and found it interesting and useful though imperfect

50 of 59

Before you invite students to use AI: AI literacy microlessons

​

51 of 59

Giving students AI context: Background materials on AI capacities, risks, and ethical concerns you might consider assigning and discussing before activities involving AI

​

52 of 59

TikTok video on ChatGPT as predictive text (3 minutes, by @mor10webn)

Share a very short video with students?�

53 of 59

Is there bias in AI treatment of your subject? An example for a French class

Prompt: Could you generate an image representing the diversity of Francophone peoples?

​

ChatGPT using Dall-E, November 8, 2023

54 of 59

Dall-E’s “Senegalese writer” vs “French writer”

55 of 59

Places to look for more ideas

​

56 of 59

See all the prompt ideas from Teaching with AI: A Practical Guide to a New Era of Human Learning by José Antonio Bowen and C. Edward Watson. Jose Bowen has made the prompts and descriptions of AI tools free on his website.

57 of 59

Ethan and Lillach Mollick’s paper “Assigning AI” describes creative approaches

58 of 59

More varieties of tutoring-style writing assistance: the PapyrusAI prompts

PapyrusAI, from the Digital Learning Lab at UC Irvine, is also funded by the California Education Learning Lab and led by Tamara Tate and Mark Warschauer.

They have shared an open library of tested feedback prompts organized by student need, such as brainstorming, organization, evidence, thesis refinement, and grammar and mechanics.

59 of 59

Questions or comments?

Good luck and enjoy as you explore!

​

�

AnnaRMills.com

LinkedIn

armills@marin.edu

​

Slides, open for commenting:

https://link.annarmills.com/FAST2�

This presentation is shared under a CC BY NC 4.0 license.