Artificial Intelligence: Readings and Resources

AI-Aware Teaching and Assignment Design

Incorporating AI

Our students report that they desire to become effective users of AI tools, but that they need guidance on ways to utilize these technologies productively in their work. Here are some constructive ways you might begin to incorporate AI into course assignments.

Ask students to analyze or critique AI-generated written content, such as an argumentative essay, a snippet of code, a mathematical proof, or a summary of an article, book, or video.
Ask students to use AI-generated visual material as an accompaniment to an in-class presentation, and to discuss or reflect on the process of creating that material.
Collaborate with students on engineering useful prompts for AI that are relevant to your subject matter, and compare and contrast the results from different AI tools.
Have your students use AI to generate outlines for writing assignments. Subject that outline to an in-class critique before students write a full draft.
Define a process that is relevant to experiential learning in your field (conversation, experimentation, etc). Then, ask an AI tool to roleplay that process with you and your students.
Ask your students for their perspective! They may have thoughts on how you can use AI in the classroom in a constructive manner.

Note: Many AI tools require users to create accounts before accessing their services, and not all AI tools meet our institutional accessibility standards. Please consider the privacy implications of requiring students to create accounts as well as any potential accessibility limitations before requiring the use of any AI tools. For more information, see the Tool Recommendations section of this guide.

Disincentivizing AI Use

MHC students value authenticity in their academic work, and are aware that relying upon artificial intelligence may, in some cases, interfere with their self-expression and scholarly reputation. However, it is not always clear to them when common applications they might use (including word processing and editing tools) have incorporated AI features. Given the potential for confusion, there are a variety of strategies we recommend you employ to discourage students from turning to AI, when you believe that its use is not appropriate.

Check the clarity of your assignment instructions. Include information on what kinds of assistance, including AI, are allowed on each assignment. If AI is disallowed, explain why; doing so helps students place the work of an assignment in the context of your goals for them as learners.
Students naturally turn away from AI when they know it will not produce work of an acceptable quality and that that lack of quality will impact their academic outcomes. You can demonstrate the limitations of these tools in class, by asking AI to respond to an assignment prompt (including citations). Then discuss how the AI result fails to meet your grading standards.
- If you find that the AI results would be passable work, consider adding either topical specificity or opportunities for metacognition to the assignment instructions, such as reflection on in-class conversation (see below).
Develop writing assignments that extend the classroom conversation. Ask students to respond to ideas and themes that emerged in discussion, or to apply your course content to a new text, item, or idea.
Restructure a vulnerable assignment to include more documented steps on the way to completion. You may choose to:
- Require a revise-and-resubmit of the written assignment of their choice at the end of the semester, as part of your final assessment.
- Incorporate opportunities for peer review of both early-stage and late-stage assignments, including outlines, drafts, and revised content.
- Ask students to submit a plan for how they intend to complete a major assignment in your course, including how long they think it will take, where and when they will do the work, and what sources of help they plan to seek out. You could require this prior to assignment submission, or as a reflective component after the assignment is complete.

Artificial Intelligence Tools (Oregon State University ECampus)

How to Create Compelling Writing Assignments in a ChatGPT Age (James Lang for the Chronicle of Higher Education)

From AI to A+: Prepare Your Students for Using ChatGPT and Other AI (Ryan Watkins, George Washington Univ.)

Teaching and Learning in the Age of genAI: Policy and Practice (Gettysburg College)

Teaching with ChatGPT: Assignment Design Tips & Ideas (Montclair State Univ.)

AI Text Generators: Resource Page (George Mason Univ.)

Concerns About AI

Turnitin’s AI Detector: Higher-Than-Expected False Positives | Inside Higher Ed - June 2023

Language Models are Few-Shot Learners | on arxiv.org - v.4, July 2020

The Carbon Footprint of Machine Learning Training Will Plateau, Then Shrink | on arxiv.org - April 2022

Sustainable AI: Environmental Implications, Challenges and Opportunities | on arxiv.org - v.2, January 2022

On the Dangers of Stochastic Parrots: Can Language Models Be Too Big? | March 2021

This paper by Emily M. Bender, Timnit Gebru, Angelina McMillan-Major, and Margaret Mitchell was published long before the current intense conversation about AI, but is worth a read, because it lays out concerns about the very development of this type of model and what it can be used for.

The Carbon Footprint of ChatGPT | Towards Data Science - December 2022

OpenAI Used Kenyan Workers on Less Than $2 Per Hour: Exclusive | Time - January 2023

Teaching AI Ethics

This post covers the questions of AI ethics from the simple to the complex, and can be adapted to students or colleagues at numerous levels of understanding.

ChatGPT: US lawyer admits using AI for case research