Monday, November 13, 2023

Instructors across campus have been grappling with how to effectively respond to generative AI (genAI) tools, and it is evident that there is no one-size-fits-all solution. Different contexts impose different kinds of constraints on the choices we can make: What might work in a small seminar might not in a large lecture course, and what novice students might be prepared to do will likely be different from advanced students who have already mastered many of the fundamentals of a discipline. GenAI tools can enhance learning, and we must also take care to ensure that students aren’t asked to share personally identifiable information on the web. Although we need to work to prevent any of the potential misuses of genAI, Dr. Amy Kristof-Brown, dean of the Tippie College of Business, argues that educating students for the world they will enter as professionals also “means thinking beyond guardrails like cheat-proof assignments and digging into AI’s real power to increase the skill sets and employability of our graduates.”  

This article outlines some broad strategies from campus leaders, instructors, and generative AI experts that you can use to design assignments for the future. 

Get to the heart of the purpose of your assignment 

Consider your learning goals as the first step in your design process. If you can articulate with some specificity what skills or knowledge you want students to develop and demonstrate by completing the assignment, you will be better able to determine if and how to alter or design your assignment to maximize student learning. Teaching expert Derek Bruff offers a framework for this thought process that helps you to identify places where genAI might interfere with or augment the purpose of your assignments and make choices for redesigning them.  

This thinking process can aid you in writing targeted genAI course policies or assignment guidelines that align with your specific context and learning goals. For inspiration, check out these example genAI syllabus statements with a number of orientations toward genAI. Note that before requiring students to use an AI tool, you must ensure that the tool is accessible for all students in your class.  

Make your assignment more specific 

GenAI outputs are limited by the content of the data they were trained on. This means that they will not be able to thoroughly complete assignments that rely on specific information that the genAI application cannot access, such as obscure texts or events that happen in the classroom. The more specific, authentic, and grounded in the context of the course the assignment is, the more impactful it will be, and the less an AI tool will be able to provide contextually useful outputs.  

You can make your assignment more specific by incorporating or combining material from the course in ways that are difficult for genAI applications to address. For example, you can ask students to: 

  • Cite in-class discussions, activities, or notes produced by the class.  

  • Cite or engage with content that is not easily accessed by genAI applications because they are relatively obscure or span across an archive.  

  • Produce work at field-specific standards

  • Apply concepts or ideas they have encountered in the course to a unique scenario. This strategy has the opportunity to produce creative assignments. For example, students can write from the perspective of a literary character, historical figure, or even non-human entity to demonstrate an understanding of course concepts.  

  • Connect ideas, concepts, or skills they learned in other courses to their work in this course (Anderson et al., 2016). 

  • Orally defend or present their assignment to demonstrate their knowledge of the material. This strategy can be added to many kinds of assignments as a way to incorporate process reflection into an assignment. 

Make your assignment more authentic 

Authentic assessments give students the opportunity to apply their learning to real-world contexts, tasks, and audiences, and be assessed according to field-specific standards (Gulikers et al., 2004). Authentic assessments can support student motivation by connecting what students are learning in the classroom to the skills and knowledge they will need to use in other contexts. As genAI continues to penetrate more industries and professional practices, some authentic assignments may center engagement and skill development with genAI. 

Authentic assignments often incorporate multimedia and multimodal elements such as presentations, slideshows, audio/visual recordings, maps, or posters, which can discourage the misuse of genAI applications.  

Some examples of authentic assignments with or without genAI include: 

  • Create for a real audience: executive summary, product proposal, policy recommendations, editorials, podcast, conference poster, blog article, patient care plan, brochure 

  • Design or create a product: diagrams, models, infographics, public-facing resource, survey 

  • Performance: role-playing, conference presentation, recorded demonstration of a skill or concept for a popular audience, interview 

Students may be authorized to use genAI to produce elements of authentic assignments without sacrificing essential learning. For example, students could use genAI to produce data visualizations as a component of a policy proposal, but their proposal would need to demonstrate a sophisticated understanding of the course material and the policy context to be fully successful. 

Emphasize and value the process 

Assignments are often used to assess a student’s learning by asking them to create a product to assess. Although genAI outputs can sometimes mimic the product of thinking and creative processes, the process rather than the product is ultimately what leads to learning. Giving students opportunities to demonstrate and engage thoughtfully with their process can disincentivize misuse of genAI as well as support thoughtful and critical engagement with genAI tools. 

Incorporating reflective components into assignments is one useful way to help students engage with their own thinking and learning process as well as demonstrate that process to you. 

Some ways to incorporate process into assignments include: 

  • Ask students to provide chat transcripts of any genAI used to complete the assignment. You can pair this strategy with a reflective component that requires students to describe how they used genAI and how it aided their process. 

  • Ask students to submit an annotated portfolio of drafts or a version history for their assignment that delineates their drafting process. This strategy can help students demonstrate that they produced the work they submitted, and it can be paired with a reflective or presentation component in which they describe how they achieved their final product and the choices they made to produce it. 

  • Scaffold the process by designing assignments with tasks that build on each other such as brainstorming, producing annotated bibliographies or literature reviews, drafting proposals, peer review, defense interviews, and final drafts. Incorporating a scaffolded process into the assignment structure will make it more difficult for students to rely on genAI to meet the assignment criteria.  

Interested in learning more? Reach out to us to schedule a consultation at teaching@uiowa.edu and stay tuned for a campus-wide Center for Teaching workshop on Designing Authentic Assignments in the Age of AI coming this spring.  


Citations

Anderson, P., Anson, C. M., Gonyea, R. M., & Paine, C. (2016). How to create high-impact writing assignments that enhance learning and development and reinvigorate WAC/WID programs: What almost 72,000 undergraduates taught us. Across the Disciplines, 13(4), 1–18. https://doi.org/10.37514/atd-j.2016.13.4.13  

Bruff, D.. (2023, July 19). Assignment makeovers in the AI age: Essay edition. Agile Learning. https://derekbruff.org/?p=4105  

Bruff, D. (2023, July 13). Assignment makeovers in the AI age: Reading response edition. Agile Learning. https://derekbruff.org/?p=4083  

Center for Teaching (2022). Authentic Assessments. Handbook for Teaching Excellence. University of Iowa – Center for Teaching. Retrieved from https://pressbooks.uiowa.edu/teaching-handbook/chapter/authentic-assessments/ .  

Center for Teaching (2023). Scaffolding. Exploring and Applying Universal Design for Learning: A Reflective Teaching Guide. University of Iowa – Center for Teaching. Retrieved from https://pressbooks.uiowa.edu/udl/front-matter/intro/

Clay, G. (2023, August 14). Why we failed to “plagiarize” an economics project with ai. AutomatED. https://automated.beehiiv.com/p/failed-plagiarize-economics-project-ai  

Clay, G. (2023a, August 14). Ai-Immunity Challenge: Lessons from a clinical research exam. AutomatED. https://automated.beehiiv.com/p/aiimmunity-challenge-lessons-clinical-research-exam  

Gulikers, J. T., Bastiaens, T. J., & Kirschner, P. A. (2004). A five-dimensional framework for authentic assessment. Educational technology research and development, 52(3), 67-86. 

Kristof-Brown, A. (2023, October 27). Teach college students to use AI proficiently (opinion). Inside Higher Ed | Higher Education News, Events and Jobs. https://www.insidehighered.com/opinion/views/2023/10/27/teach-college-students-use-ai-proficiently-opinion .  

Messier, N. (2022). “Authentic Assessments.“ Center for the Advancement of Teaching Excellence at the University of Illinois Chicago. Retrieved November 10, 2023 from https://teaching.uic.edu/resources/teaching-guides/assessment-grading-practices/authentic-assessments/