Three Use Cases for AI in the Classroom
The rapid advancements in large language models (LLMs) are prompting me to rethink my approach to teaching and learning. As I adapt, I’ve identified three distinct roles for AI in my classroom.
1. Course Design
Before the class even begins, AI can assist in course design. While I often find AI-generated course structures to be uninspired, experimenting with supplying the LLM my own writing on the content topic has yielded better results. Currently, I use AI suggestions as a starting point, refining them to create a more engaging and personalized course structure.
2. Creating Quizzes and Assignments
AI excels in handling the otherwise tedious tasks of creating quizzes and assignments. By feeding the AI examples of my related content, it generates assignments and useful grading rubrics tailored to my teaching style. This offloads some of the least desirable tasks, allowing me to focus on more critical aspects of teaching.
In the fall, I plan to experiment with a third-party grading system. Meanwhile, I've been using AI to build my own grading assistant. By providing the AI with my content and grading rubrics, it can perform an initial review of anonymized papers. This has proven helpful in suggesting areas where I can offer better guidance to students.
While this process is currently tedious and time-consuming, third-party providers are making significant strides in AI-facilitated grading. This technology can offer faster, more efficient grading, allowing for multiple rounds of feedback before final submission. This is a win-win, benefiting both students and instructors.
3. One-on-One Student Tutoring
The most exciting application of AI is in one-on-one student tutoring. Early experiments were problematic, with AI getting easily confused by already confused students. However, as LLMs improve, so does their ability to provide accurate assistance.
In the fall, I plan to provide students with prompts and instructions to direct the AI to related course content. As students work through the material, the AI can offer on-the-spot clarifications. Research consistently shows that one-on-one instruction is superior to lectures, as it addresses individual learning paces and knowledge gaps. Virtual tutors, especially when paired with a flipped classroom model, can eliminate the dilemma of either slowing down for less-prepared students or pressing ahead for those better prepared, leading to higher quality and more uniform outcomes.