As generative artificial intelligence (AI) becomes increasingly ubiquitous in higher education and here at SCU, we’re all navigating how to guide students in its responsible use. Faculty take many different approaches to AI in the classroom, from fully embracing it as co-author to prohibiting all use, and everything in between.
As with any teaching tool, generative AI delivers both benefits and challenges. The latter includes issues of privacy, racism, and sexism (information is scoured from the internet after all), misinformation, fake citations, and even malicious content. Decades of research on academic integrity (apart from the use of AI) directs us away from an exclusively punitive approach and toward pedagogical practices that acknowledge what academic integrity looks like in your discipline.
What uses of AI you allow in your classes is your decision as an instructor; whatever you decide, it’s important to be clear with your students about your expectations, and to help them understand your rationale for your policies and practices around AI. Below are some strategies and considerations for creating a transparent culture of AI use in your classroom.
Syllabus Statements
Many faculty members have already begun incorporating syllabus statements about using AI critically and ethically, such as the example below:
Example syllabus statement 1 | Example syllabus statement 2 |
---|---|
It’s important that you:
|
In this class, your unique voice—shaped by your cultural and linguistic background—is central to your writing. As many of you are first-generation, multilingual, and BIPOC students, we recognize the richness that your experiences bring. To support your development as writers, we will use AI tools like ChatGPT thoughtfully and with purpose, keeping in mind the principle of linguistic justice. Linguistic Justice and AI ToolsLinguistic justice values all forms of expression, and in this class, your voice matters. AI tools, which are built on mainstream and often white-dominant perspectives, can be useful, but they don’t always capture diverse voices or experiences. Our goal is to use AI to enhance your work, not to erase your authentic voice. |
These statements help students to understand your rationale for allowing or prohibiting certain uses of AI in your course. Doing so allows them to learn more about what AI does well, and what limits it has. Such a statement is also best used as the beginning of a conversation with your students about these issues, in which you might further explore the benefits and limitations of AI technologies for your subject area, perhaps developing your own class compact for its use.
Citing AI As a Source
If you allow the use of generative AI in your course, whatever your policy is, it is essential to clarify how your students can transparently cite their use of AI. We’ve generated a few ideas about potential student uses of AI and examples of how they might cite AI appropriately. The examples below were generated interactively with ChatGPT 4.0 and subsequently edited for clarity. You may also be interested in teaching students how to formally cite AI in different styles, and guidelines for APA, MLA, and Chicago styles are included below.
1. Research assistance
Potential uses: summarizing articles, finding sources, or explaining complex topics.
Example attribution: "I used ChatGPT (OpenAI, 2024) to help identify potential research articles related to urban heat island mitigation, which I then reviewed and incorporated into my literature review."
2. Data analysis
Potential uses: identifying patterns in datasets.
Example attribution: "AI tools such as ChatGPT (OpenAI, 2024) assisted with generating initial R code to analyze the YRBUILT variable in the AHS_NF_23 dataset, which I modified for accuracy."
3. Proofreading and grammar check
Potential uses: catching errors in grammar, punctuation, and style.
Example attribution: "Grammarly (2024) was used to review the final draft of this paper for grammatical accuracy and clarity in the writing style."
4. Generating ideas for projects
Potential uses: providing new and creative ideas for research projects, presentations, or assignments.
Example attribution: "ChatGPT (OpenAI, 2024) provided brainstorming suggestions for potential project directions, including the idea of exploring climate-sensitive mortality outcomes."
5. Studying and test preparation
Potential uses: creating customized study plans based on a student’s learning style or areas of difficulty.
Example attribution: "AI-generated flashcards (2024) helped me prepare for the midterm by focusing on key terms and concepts from the course readings."
6. Summarizing class notes
Potential uses: summarizing or organizing notes from lectures or reading materials.
Example attribution: "I utilized Otter.ai (2024) to transcribe and summarize class discussions, which helped inform the reflections presented in this ethnographic analysis."
Still New to AI Yourself?
Below are a few insights that might be helpful to consider when you’re thinking about AI in your classroom:
1) AI Tools Generate Information (but they don't analyze it, it may not be accurate, and they make stuff up)
AI tools like ChatGPT are large language models (LLMs) that scour the internet for information and make connections between words and phrases. ChatGPT, in particular, has been used to complete writing tasks, generate Excel-related prompts, draft lab reports, and do coding assignments quickly. They often lack the sophistication, details, critical perspective, and specificity to course content and contexts that we expect in excellent student writing. For example, it can explain theoretical concepts well (because of an abundance of texts on the topic online), but it cannot generate a deeper analysis of the theory. It also cannot draw information from non-text modalities like videos, nor can it, obviously, draw from non-internet sources, such as the nuanced and particular discussions that you and your students are having in class. Oftentimes, ChatGPT-generated writing is recognizable because of its use of repetitive phrases or stylistic constructions, and a lack of specific contexts. It’s worth talking with students about what ChatGPT can and cannot do: students are still learning about the affordances and limits of these technologies. So are we. So is everyone.
2) AI and Digital Pedagogy
When you are inclined to worry that the world of knowledge (and teaching and learning) is coming to an end, think about the use of new AI tools as an extension of already existing digital pedagogical practices. Digital pedagogy can look different depending on the instructor and their discipline: lectures may be paired with online research conducted by students to create a digital artifact. Faculty may use the “Designer” feature on PowerPoint to make their presentations more visually appealing without a lot of work. Faculty may allow or even encourage students to use Grammarly and other similar tools to check their writing. Faculty may also use such tools themselves. With the abundance of available digital tools faculty can set boundaries and expectations by talking to students about their use of generative AI tools. How do you expect students to use these tools (if at all)? Why? Can you imagine using generative AI tools to support student learning? Many faculty here and elsewhere are addressing these issues in syllabus statements and assignment prompts. As with many technological tools, talk to your students. Don’t assume they know what’s expected or appropriate.
3) Focus on Increasing Access and Equity
ChatGPT and similar AI tools can increase access to learning resources for students with diverse learning needs. Captioning, audio description, text-to-speech, and speech-to-text are examples of generative AI tools that support inclusive learning. Additionally, students who struggle with writing may benefit from using ChatGPT to help them develop their ideas or improve their writing. Collaboratively writing with ChatGPT, for example, may provide an initial structure for students to strengthen their positions, enhance and specify their arguments, integrate their own voice and perspective, or edit for clarity. Generative AI may help students reduce stress and spend more time engaging with course content to deepen their learning. Although Generative AI, including ChatGPT, may offer important assists to student learning, it cannot offer personalized support, mentorship, and the relational aspect of teaching and learning, which is fundamental to what we all do at SCU.
Acknowledgements and Additional Resources
The Faculty Collaborative for Teaching Innovation would like to thank our SCU colleagues and speakers at our AI focused talks, events, and scholarship:
- Eric Haynie (Manager, Instructional Technology)
- CJ Gabbe (Teaching Professor, Environmental Studies & Sciences; Faculty Associate)
- Cynthia Alby (Professor, Teacher Education, Georgia College & State University)
- Matt Gomes (Teaching Professor, English, SCU)
- Denise Krane (Lecturer, English & Director, HUB Writing Center)
- Maya Ackerman (Teaching Professor, Computer Science & Engineering)
- Melissa Brown (Teaching Professor, Communication)
For more resources about teaching with AI, check out the following resources:
SCU’s Repository of ChatGPT Resources
Check out SCU’s Markkula Center’s Hackworth Fellows student-generated discussion and guidelines for the ethical use of AI.
Classroom Policies for Using AI Generative Tools (Georgetown University)
Page authors: Lisa Chang, Eric Haynie, CJ Gabbe
Last updated: October 7, 2024