TLI, along with several other offices and units across the university, recognizes the growing interest in artificial intelligence and the many consequential impacts it might have on teaching and learning on all levels. Therefore, we have assembled this webpage as a resource for faculty members, graduate students, and members of our campus community.
Artificial intelligence capabilities and availability are changing rapidly, and so are experts’ approaches to their best use in higher education. On this page we offer a current snapshot of an ongoing conversation between instructors, students, and a variety of professionals with expertise in this area. As the conversation continues, we will make updates to this webpage. You are invited to return here regularly to access the latest ideas on how to approach AI as it relates to teaching and learning.
If you would like to get involved in these conversations, we also invite you to consider joining our Learning Community on Artificial Intelligence in the Classroom, where you can talk with faculty and staff who are also committed to enhancing the university’s response within and beyond the classroom.
Basic Introductions to Artificial Intelligence Writing Tools
- Educator Considerations for ChatGPT – OpenAI, the company that created ChatGPT, has posted this resource with basic information about what ChatGPT is, and some education-related risks and opportunities.
- Will Douglas Heaven, “The Inside Story of How ChatGPT Was Built from the People who Made It” — an interview with OpenAI staff, published in MIT Technology Review.
- Cynthia Alby, “ChatGPT: A Must-See Before the Semester Begins” – Cynthia Alby, a Professor of Teacher Education, provides a basic explanation of OpenAI’s new AI Writing Tool, ChatGPT, and why instructors should be prepared to respond to it.
- Emma Bowman, “A new AI chatbot might do your homework for you. But it’s still not an A+ student” – In this article, NPR reporter Emma Bowman provides a helpful summary of what ChatGPT can and cannot do.
- Meghna Chakrabarti, “How AI chatbots are changing how we write and who we trust” – In this podcast episode of NPR’s On Point, Meghna Chakrabarti discusses what AI writing can and cannot do and how it might impact educators.
- College Unbound, “AI Generative Tools Policy Development Plan” – College Unbound has created a helpful primer on ChatGPT, as well as a model of the kinds of policies an instructor or institution might implement.
- Beth McMurtrie, “Teaching Experts Are Worried About ChatGPT, but Not for the Reasons You Think” – Beth McMurtrie, Senior Writer at the Chronicle for Higher Education, explains OpenAI’s ChatGPT and why instructors are worried about it.
- OpenAI, “ChatGPT: Optimizing Language Models for Dialogue” – in this blog post, OpenAI, the company that designed the popular AI writing tool ChatGPT, describes how the technology works.
- “Chat GPT Writing & Science Demo” – by Michael Palmer, Center for Teaching Excellence, University of Virginia. In this video, Palmer shows how ChatGPT can draft course materials and creative texts based on basic prompts.
How to Identify Writing Composed by Artificial Intelligence
- Emma Bowman, “A college student created an app that can tell whether AI wrote an essay” – This NPR article explains Edward Tian’s GPTZero: an app that aims to catch uses of ChatGPT.
- Chris Carin, “AI writing: The challenge and opportunity in front of education now” – Chris Carin, CEO of plagiarism detection company Turnitin, explains that Turnitin is does have some AI detection tools and is planning to roll out more in 2023.
Compilations of Additional Resources
- Writing Across the Curriculum Clearinghouse, “Starting Points for Inquiry,” a resource compiled by the clearinghouse’s editors, including shared resources and opportunities to get involved.
- Lance Eaton, Classroom Policies for AI Generative Tools — a Google wiki document where you and other faculty members can submit your classroom policies on AI. This is a public document, edited by a wide range of faculty members, and does not carry the endorsement of any institution (added 1/27/23).
- Anna Mills, for the Writing Across the Curriculum Clearinghouse, Sources to Stimulate Discussion among Teachers — a list of sources across a wide variety of areas of inquiry and instruction (added 1/27/23).
- Lee Skallerup Bessette, “ChatGPT” – A Zotero library full of resources about ChatGPT, compiled by Lee Skallerup Bessette, Assistant Director for Digital Learning at Georgetown University.
- SUNY Online, “ChatGPT – Learn More” – A collection of resources compiled and a sample syllabus statement about the use of AI writing tools in the classroom
- Zehnder et al, “Breaking News: ChatGPT Breaks Higher Ed” – Caralyn Zehnder, Cynthia Alby, Karynne L M Kleine, and Julia Metzker have created a helpful set of curated resources related to ChatGPT.
Inclusive Teaching
- As stated in the Higher Education Opportunity Act (2008), one scientifically valid framework for guiding educational practice to reduce barriers to instruction is Universal Design for Learning (UDL). One of the key principles of UDL is flexibility in the ways students respond or demonstrate knowledge and skills. The emergence of AI writing tools provides an opportunity to integrate different types of assignments into your classroom, which could allow students to demonstrate their learning in new ways. For example, AI writing tools are currently incapable of creating podcasts, videos, or images, or conducting an interview or giving a presentation in a class. Some students might relish the opportunity to demonstrate their knowledge in an alternative way.
- Students thrive in “learning environments that students perceive to be inclusive and affirming and where expectations for performance are clearly communicated and set at reasonably high levels” (Pascarella, 2001; Kuh, 2001). Classrooms that take student’s individual differences into account and are learner-centered also increase student engagement and participation (Hockings et al, 2008). As you make changes to address AI writing, be sure that your classroom remains an inclusive space, and that you are not unintentionally adding barriers that affect students in inequitable ways.
Goals & Motivation
- Clearly articulating expectations so that students know the desired outcomes helps students understand the actions they need to take to meet that outcome, creating a more positive outcome expectancy (Ambrose et al, 2010). If you expect your students to use or not use AI writing in your classroom, be sure to clearly articulate this to your students (e.g., via a paragraph on your syllabus) and explain why. Students may be more likely to meet your expectations if they understand your reasoning.
- Honor codes are a research-based means of reducing and hindering academic integrity issues from starting (McCabe, 2012; Lang, 2013). To help discourage students from using AI writing tools in ways you have prohibited for your class, discuss the University of Tennessee Student Code of Conduct with your students.
- Course-specific learning goals help to guide students’ activities (Simon and Taylor, 2009). Consider whether your learning goals could be met with the addition of assignments that do not solely rely on writing, such as multimedia assignments, interviews, or peer instruction. Alternatively, consider whether the use of AI writing tools might actually help your students meet your learning goals more effectively, if integrated into the course.
- Students value course goals and activities more when they are connected to students’ interests, provide authentic real-world tasks, and are relevant to their current academic lives (Ambrose et al, 2010). Coincidentally, AI writing is unable to answer questions about students’ personal lives or about current events that may be affecting them. Consider designing assignments that ask students to connect course content to their personal lives or that are connected to events that have occurred since 2021 (the end of the GPT-3 training data that many AI writing tools use as their basis). Also remember that AI writing tools themselves are now relevant to your students. Having discussions about this new technology and how they might use it may foster student interest.
- Students who are not motivated are more likely to take shortcuts (Lang, 2013). Instructors can make use of the TARGET strategy to increase student motivation, increasing the chance that students only use ChatGPT in ways that support, rather than supplant, their learning.
- Targeted feedback enhances the quality of students’ learning (Ambrose et al, 2010). It also has the benefit of demonstrating for students that you are, in fact, carefully reading what they turn in. Evidence of your investment may make students less likely to use AI writing tools in ways that decrease their opportunities to learn.
- Tools such as PowerNotes, supported by the UTK Libraries, help students track and record information from online sources, retaining URL information automatically and allowing students to focus on and engage with the content.
Effective Course Design & Active Learning
- Becky Supiano, “Will ChatGPT Change How Professors Assess Learning?” — in this essay, Supiano reviews basic information about how ChatGPT works, and how it might influence assessment methods.
- Infrequent, high-stakes assignments amplify pressures on students, and may lead them to attempt shortcuts (Lang, 2013), whereas breaking an assignment down into a series of lower-stakes assignments provides more opportunities for practice, which improves student learning (Ambrose et al, 2010). Instead of assigning one high stakes writing assignment, break the assignment down into various, lower-stakes, scaffolded assignments. AI writing tools will show varying results for these kinds of assignments, in ways that may help students better understand the tools’ limitations.
- The flipped classroom model, in which the typical lecture and homework elements of a course are reversed, is based on the evidence-based strategy of promoting active learning (Freeman et al, 2014). Flipped classrooms also allow assignments to be completed in the classroom, where students may be less likely to use AI writing to take the place of their own learning.
- Providing a model of a successfully completed assignment can help to clarify what students need to do. Providing a model of what you do not want can also be helpful (Ambrose et al, 2010). Consider using ChatGPT to provide examples of completed assignments that you can evaluate with students, to help derive a consensus about the tool’s abilities and limitations.
- Research shows that graphic organizers such as concept maps, diagrams, charts, graphs, and grids help students organize ideas, represent relationships, and retain information (Narkawicz & Casteel, 2012). These are also assignments that AI writing tools cannot easily produce. Consider bringing more of these assignments into your classroom.
- Collaborative learning and group work are beneficial for student learning (Hassanien, 2005). To take advantage of this and make it more difficult for students to use ChatGPT in ways that may limit their learning, consider assigning collaborative writing projects (e.g., asking students to work together on a Google Doc).
Bringing Artificial Intelligence Tools Into the Classroom
- Becky Supiano, “Will ChatGPT Change How Professors Assess Learning?” — in this essay, Supiano reviews basic information about how ChatGPT works, and how it might influence assessment methods.
- Cynthia Alby, “Can ChatGPT Be a Blessing?” – In this essay, Cynthia Alby,a Professor of Teacher Education, argues that instructors should pay attention to the positive uses students might put ChatGPT to, instead of worrying about cheating.
- Autumn Caines, “ChatGPT and Good Intentions in Higher Ed” – In this blog, writer and lecturer Autumn Caines discusses some of the problems that an instructor should consider before integrating AI writing into their classroom, such as concerns about how student data might be used.
- Ethan Mollick, “The Mechanical Professor” – In this blog post, Dr. Ethan Mollick illustrates how an AI-powered chatbot can approximate much of the work he does as an Associate Professor of Management, including drafting syllabi, writing lectures, grading assignments using a rubric, creating assignments, writing an academic paper, and writing code to reshape a data set.
- Andrew Myers, “CoAuthor: Stanford experiments with human-AI collaborative writing” – This article (written by Andrew Myers, a contributing writer for the Stanford Institute for Human-Centered AI) reflects on the potentials and possible challenges raised by CoAuthor, a large language model created by Mina Lee and Percy Liang and based on OpenAI’s GPT-3 model. CoAuthor’s creators study of 60 subjects who used the technology to write found that works produced by this ‘partnership’ had fewer spelling and grammatical errors, writers were more productive (by word count), and that the collaboration had little impact on writers satisfaction with the writing process. However, it could have a negative impact on sense of ownership of the created texts.
- Marc Watkins, “AI Will Augment, Not Replace” – In this opinion piece in Inside Higher Ed, Marc Watkins, Lecturer in Composition & Rhetoric at the University of Mississippi, argues that students must be taught how to use AI writing tools and describes work he did in an AI working group, using the technology to help students brainstorm research topics.
For more information for instructors about how to handle A.I. capabilities in courses that use writing assignments, please see the Judith Anderson Herbert Writing Center’s resource page, AI Tools and Writing Assignments: Instructor Resources.