Strategies for Effective Teaching in the Age of AI

AI’s opportunities for teaching and learning are as vast as the concerns it raises about assessing it. Whether or not you buy into the “homework apocalypse” rhetoric, AI is challenging us as instructors to refresh our pedagogy and provide new ways for students to demonstrate their skills, knowledge, and abilities. You can use the strategies below to create more effective learning experiences, both AI-enabled and not.

Promote metacognition.

Regardless of whether students are using AI to complete assignments, incorporating tasks that ask students to reflect on and analyze their own learning creates an opportunity for them to demonstrate competence and enhances overall success. Helping students develop a critical awareness of their learning improves their ability to transfer this self-knowledge to other domains.

With AI: You might ask students to track when they used AI and why. Did it help them to clarify a point that they didn’t understand? How did using the tool further or complicate their understanding of a topic? Did using AI help them meet the goals of the assignment?

Without AI: What strategies and skills did students use to complete the assignment? Were they effective? What resources did they use or need to be successful? What patterns did they notice in their process of learning? How has their attitude toward their learning and the content shifted?

Get specific.

There is no such thing as an “AI-proof” assignment. With enough elbow grease, AI can be leveraged to accomplish almost any task. Though you may not be able to prevent students from using AI in unauthorized ways, you can design assignments that they actually want to do by making assignments authentic to their experience and relevant to their lives.

• Ask students to connect assignments to specific readings from class, or use a point raised in discussion to defend or critique a stance.
• Incorporate a reflective component that asks students to explain how the assignment relates to their personal beliefs or professional aspirations.
• Explore current or local problems that AI doesn’t have access to by using examples that are local to the Notre Dame community or from recent months. Just be mindful that while ChatGPT 3.5 isn’t connected to the internet, ChatGPT-4, Microsoft Copilot, and Google Gemini are, and have access to real-time information, with mixed results. For example, when we asked Gemini to provide us with a list of citations for scholarly articles about women mystics in the Middle Ages, only one of the results was written this century and directed us to academic databases for a more thorough search. Try running your assignments through each platform to see how the AI responds before giving them to students—just be mindful of your own generative AI privacy settings, as your assignments may be collected and used as part of the model’s training data.

Emphasize process over product.

Scaffolding assignments into smaller checkpoints can deter—though not prevent—unauthorized use of generative AI. If the assignment is AI-enabled, working in steps provides the opportunity to guide students through using AI effectively and ethically. This allows you to give feedback throughout the process and see how students are developing ideas and changing their thinking over time. It also invites metacognition; a final portfolio, for example, lets students create a narrative about the work they have completed from start to finish and explain their choices.

Other ways to demonstrate process:

• outlines or proposals
• annotated bibliographies
• drafts and feedback

You might also consider having students annotate class readings, either independently or collaboratively using Perusall. Rather than relying on generative AI for summaries of articles or concepts, annotations help students process and understand complex readings and improve literacy.

Check it out: Derek Bruff, author of Intentional Tech: Principles to Guide the Use of Educational Technology in College Teaching (2019), has several blog posts chronicling his efforts to redesign assignments to accommodate AI-based disruption. His posts on reading responses and essays are full of great tips. Schedule a consultation with us to talk about redesigning one of your own assignments.

Flip the classroom.

Traditional classrooms are structured so that “first exposure” to new concepts and topics happens in class, usually through a lecture. Students then practice and apply their knowledge by completing homework assignments outside of class. Flipping the classroom, well, flips this structure. Instead, students are responsible for introducing themselves to the material through pre-recorded lectures or readings, and complete checks for understanding that prepare them to participate in class. Class time is spent applying what they’ve learned and receiving feedback, usually through active learning.

Examples of a flipped approach:

• Have students annotate readings and come prepared to discuss the main arguments and draw connections between them.
• Students watch a pre-recorded video on a concept and complete a pre-class assessment, which allows you to focus class activities on addressing gaps in knowledge.
  • Students might work together on a problem set or case study, or work in teams to solve a problem.
  • Map or draw a concept, then explain and compare their drawing with a partner.

In-class practice limits students’ ability to use AI tools in unauthorized ways while also providing an opportunity for students and instructors to experiment with how AI can be used to work on disciplinary problems.

Check for alignment.

The reality that AI can perform lower-level or routine tasks efficiently and accurately challenges us to rethink the purpose of our assignments and what kind of learning we are trying to get our students to demonstrate—in other words, are our learning goals actually what we want for our students, and are our current assignments helping them get there? Learning goals that emphasize the process of learning rather than the product of learning help us create assignments that ask students to engage in more complex and higher-order thinking. Inara Scott, a professor at Oregon State University, developed the “less content, more application” challenge to help move students away from content-based “what” knowledge and toward skills-based “when, why, and how” practice.

This is not to say that students don’t need to know the fundamentals—they certainly do! However, we should be transparent with students about why we are asking them to complete assignments and tasks that could easily be done with AI. Emphasize the importance of the skills they are developing and how they contribute to the value of their learning. Once students have gained proficiency in the basics, AI can help them continue to develop their skills through activities that ask them to analyze, critique, and revise AI-generated output.

You may also want to consider alternative or multimodal assessments that allow students to demonstrate their knowledge of fundamental concepts in non-traditional ways to encourage a more relevant and authentic engagement with the content.

Cultivate inclusivity.

In Cheating Lessons: Learning from Academic Dishonesty (2013), James Lang identifies four conditions that underlie academically dishonest behaviors: 1) an emphasis on performance, 2) infrequent high-stakes assessments, 3) low expectations of success, and 4) extrinsic motivation alone, such as grades (35).

When we create inclusive and supportive learning environments that:

• cultivate belonging for all students,
• affirm students’ identities, experiences, and abilities,
• express confidence in students’ ability to learn and succeed,
• provide multiple ways for students to demonstrate their learning, and
• are relevant to students’ lives,

we take steps to foster intrinsic motivation in students—the desire to complete a task for its own sake, either out of interest or enjoyment. The strategies outlined above that connect assignments to students’ lives and experiences contribute to belonging and motivation. Likewise, metacognitive assignments hold students responsible and accountable for their learning. Assignments that allow students to demonstrate their learning in multiple ways provide students with the opportunity to be creative, agentic, and leverage their prior skills and abilities. In inclusive classrooms, students may also feel more comfortable seeking help or further guidance when they don’t understand a concept, mitigating the recourse to academically dishonest behavior. Finally, reasonably flexible grading and late-work policies can lower the pressure and anxiety around assessments.

However, AI can also be an asset for more equitable learning. Generative tools like ChatGPT can, with proper checks, help students break down complex topics, expose students to different perspectives, brainstorm ideas, and focus. Maggie Melo, a professor at the University of North Carolina-Chapel Hill, has written about using ChatGPT as someone with ADHD. And students are using it to help them decode assignments:

 

Check it out: Ethan and Lilach Mollick have written extensively about integrating AI technologies into evidence-based teaching practices. “Using AI to Implement Effective Teaching Strategies in Classrooms” is a great place to start.