Prohibiting the use of GAI entirely is unlikely to be completely successful, just as attempts to prohibit the use of calculators and the internet were. New technologies infiltrate society quickly if they are useful, and young people are often the earliest adopters. However, if you decide that the necessary learning outcomes for your course/assignment are achievable only in the absence of GAI input, here are some quick ways that you can make GAI less useful to students in completion of the assignment, and some ways to increase compliance with the prohibition.

Making GAI Less Useful #

1. In-class activities: It is nearly impossible for students to leverage GAI tools to complete activities in a proctored or live context. Live discussions, formal debates, group concept mapping, and hands-on activities like lab exercises can be employed formatively, helping students develop their knowledge, skills, and dispositions, and providing the students with feedback to stimulate improvement. Proctored quizzes and examinations, including blue-book exams, oral exams, and presentations are useful in assessing some student proficiencies. You may find it helpful to flip your classroom—covering “content” before class-time outside of class, so you have more time during class for these types of activities.
2. Reflections: Personal reflections in which students are asked to incorporate their own experience, feelings, and opinions are difficult for GAI to generate authentically. These are most effective if they ask students to relate to course material in a unique way, or engage students in referring to previous discussions, lectures, or assignments that GAI cannot have been trained on. 
3. Peer reviews. These provide authentic opportunities to deliver feedback and learn from, and with, other students. They encourage collaboration, develop critical thinking skills, and are hard to fake with GAI, especially if the written review is paired with discussion sessions. (For GAI to be used in producing a peer review, students would have to upload somebody else’s intellectual property into a GAI tool. It would be worthwhile to have a discussion of the ethics of this and how they relate to academic integrity.)
4. Authentic assessments. Projects that are tied to real-world issues or tasks relevant to the students’ field of study increase learning and engagement, and are difficult for GAI to complete. As a strategy, encourage students to integrate concepts or methodologies they are learning in their other concurrent courses. This multidisciplinary approach fosters a more comprehensive understanding of the material and promotes the application of diverse academic skills, making it challenging for GAI to generate responses while enhancing the students’ learning experience.
5. Time-limited assignments. Providing assignments that are to be completed within a specific timeframe (while ensuring it’s reasonable for completing the work assigned) can reduce the likelihood of students relying on GAI. The rationale behind this is similar to using time-limited exams in online courses as a means of inhibiting googling all of the answers.
6. Capstone projects: Require a final project that ties together all the concepts learned in the course. The complexity and personal engagement required for a capstone project can’t be easily outsourced to GAI.
7. Student-created content: Have students create and share content relevant to the course, like a blog post, video, or podcast. GAI will struggle to create such content without obvious signs of non-human creation.
8. Group work: This is a good way to encourage collaboration and engagement. Having students work in pairs or small groups on a shared project can deter the use of GAI, as each group member can help hold the others accountable.
9. Assess the process, not just the product. This practice underscores the importance of the problem-solving process, not just the final solution itself. It also helps instructors pinpoint where misconceptions or mistakes occur and tailor feedback and support to where it is needed. By focusing on the learning process rather than products, we can make it harder for students to use GAI tools inappropriately and foster a more engaging and meaningful learning experience for all. Here are some specific strategies for assessing the processes rather than the products of learning, which can be adapted for multiple teaching modalities.
   1. Scaffolded assignments: Break down larger tasks into smaller, manageable parts, each with its own deadline. This allows you to assess understanding at each stage of the learning process.
   2. Metacognitive reflection: Have students keep a journal to document their learning process, reflect on the challenges they faced, and explain how they overcame them. Not only does this type of assignment deter cheating, it also creates a channel for students who struggle with traditional assessments to demonstrate their thinking and learning, and helps all students become more aware of their own learning process.
   3. Process portfolios: Students can create portfolios that document their progress over time, showing how their understanding of a topic has developed.
   4. Follow the bread crumbs: Encourage students to document key insights from class discussions, collaborative group tasks, and solo activities that occur within the classroom setting. Incorporate these gathered insights into a comprehensive project or final examination, demonstrating the progression and maturation of their critical thinking skills. This process not only exhibits their grasp of the course material but also provides a chronological reflection of their evolving intellectual journey.
10. Use a multi-modal assignment: Multimodality refers to the use of multiple modes or media within a single artifact. Instead of a text-centric consider integrating visual, auditory, spatial, and/or tactile elements. This encourages students to think more creatively and critically about how they present their knowledge and offers students multiple ways to demonstrate their understanding.

Increasing Compliance #

1. Integrate academic integrity education into your course. Establish a course-wide framework for integrating academic integrity instruction. This is vital to obtaining student valuation for following the rules that you establish in your course for GAI use. Consult CITL resources for integrating academic integrity in your teaching.
2. TILT: Transparency in Learning and Teaching goes a long way toward increasing student valuation of assignments. The more students value assignments, the less likely they are to take shortcuts or to cheat. Making it clear in the assignment introduction what the purpose of the assignment is, what students will gain (knowledge/skills) from completing the assignment, and how the assignment will prepare students for careers or future coursework, provides the context students need to invest time and effort in the assignment. You should also be clear, in advance, how you intend to evaluate the assignment, with explicit guidelines, expectations, and scoring criteria or a detailed rubric. Taken together, these help students to better define the target they are attempting to hit which can provide them with more confidence in their ability to complete the assignment. You can download teaching tools for making your assignments more transparent from Tilt Higher Ed.
3. Ungrading: A grade-oriented (as opposed to a learning-oriented) approach often motivates students to cheat and rewards them for doing so. Alternative assessment approaches, such as specifications grading or ungrading, introduce a paradigm shift that encourages faculty to design courses in a way that rewards exploration and growth through meaningful effort rather than as a series of compulsory, grade-focused tasks. On a small scale, faculty can experiment with offering weekly ungraded quizzes. This practice provides a low-stakes opportunity for students to gauge their understanding and reflect on their learning process. For faculty, it provides feedback on what students are learning, where they struggle, and data on individual performance and growth over the semester.
4. Collaborative assignment design: Collaborating with students in the design or grading specifications for an assignment produces greater student ownership of their education.

Adapted from Teaching with Artificial Intelligence, Presented by the Biggio Center for the Enhancement of Teaching & Learning at Auburn University Module 5 Support Student Use of AI. 

References and Resources: #

Biggio Center for Enhancement of Teaching & Learning at Auburn University  Teaching with Artificial Intelligence, Module 5 Support Student Use of AI.

Bowen. AJ. And C.E. Watson, (2024) Teaching with AI: A Practical Guide to a New Era of Human Learning.  Johns Hopkins University Press, Baltimore.

McDonald, N., Johri, A., Ali, A., & Hingle, A. (2024). Generative artificial intelligence in higher education: Evidence from an analysis of institutional policies and guidelines (arXiv:2402.01659). arXiv. 

Cengage, (2023) 8 Ways to Prevent Students From Cheating With AI

Kelly, K.J. 2023. Teaching Actual Student Writing in an AI World, Inside Higher Education, January 18, 2023.