Jessica Bernards is a math professor at Portland Community College and University of Oregon. This instructional example was designed for a second-semester college calculus course but the structure of the activity is intentionally transferable and can be adapted to any course or subject area. Students use a generative AI tool as a tutor, but only after first explaining the concept back to the AI in their own words. The AI evaluates the student’s explanation, prompts refinement when needed, and then provides guided practice problems where it gives feedback based on student individual answers. Students must solve problems independently, use the AI to diagnose errors in their work, and reflect on the learning process.
This activity incorporates multiple evidence-based teaching practices, including metacognition through structured reflection, and formative assessment and practice. Rather than using AI to produce answers, students use it as a process-focused learning partner to enhance their learning about integration with u-substitution and with partial fractions. The lab creates individualized learning pathways, as each student’s interaction with the generative AI is shaped by their current level of understanding, misconceptions, and pace of learning.
Digital Resources
Google Gemini or any LLM AI’s study and learn capability
The primary digital tool used is Google Gemini (but it could be any LLM AI as they all have the study and learn capability), accessed through institutional accounts. Students interact with Gemini using carefully structured prompts that frame the AI as a tutor rather than a solution engine. This tool provides immediate, adaptive feedback based on each student’s explanation, effectively functioning as a personalized tutor that meets students where they are and will never give students answers.
The digital tool supports individualized learning at scale by allowing every student to receive targeted guidance, clarification, and practice without waiting for instructor availability. This flexibility is especially valuable in online and high-enrollment courses, where one-on-one support is difficult to provide consistently.
Digital Enablement
This AI chatbot activity creates truly individualized learning experiences by meeting students at their current level of understanding and adjusting in real time based on their explanations and questions. Each student effectively designs a personalized tutor that responds to their specific gaps, misconceptions, and pace of learning. Using an LLM AI tool accessed through institutional accounts levels the playing field for students who may not have access to private tutoring, extensive outside support, or prior exposure to advanced mathematics. By providing on-demand, process-focused guidance at no cost, the activity reduces reliance on external resources and helps ensure that all students have continuous access to academic support. Framed intentionally, the AI functions as a free tutor students can carry with them, empowering independent learning while preserving instructor oversight and academic integrity.
Students demonstrate deeper conceptual understanding, improved ability to identify and correct errors, and greater persistence with difficult material. Because feedback is immediate and tailored, students are more likely to continue working through challenges rather than disengaging. Bernards and fellow instructors observe reduced reliance on memorization and stronger self-regulated learning behaviors in students who have completed the lab. Students also report increased confidence and a clearer understanding of how to use AI responsibly to support learning rather than to replace it.
This activity supports equitable outcomes by providing all students with free, on-demand access to individualized academic support, reducing disparities tied to access to private tutoring or informal academic networks. It explicitly teaches how to use AI ethically and effectively, helping demystify tools that students may otherwise use inequitably or incorrectly. The lab emphasizes process over speed, normalizes struggle, and includes reflective components that value growth rather than prior preparation. To implement equitably, instructors should provide clear framing, example reflections, and flexible submission options, ensuring the AI functions as a support tool that enhances learning while maintaining academic integrity.