Spotlight on SOTL: Using AI to Augment Teaching

This article is base don the following research article:

Mollick, Ethan R. and Mollick, Lilach, Using AI to Implement Effective Teaching Strategies in Classrooms: Five Strategies, Including Prompts (March 17, 2023). The Wharton School Research Paper, Available at SSRN: https://ssrn.com/abstract=4391243 or http://dx.doi.org/10.2139/ssrn.4391243

This paper provides practical guidance for instructors on how to leverage AI tools like ChatGPT to more easily put into practice teaching techniques that research has shown to be effective for student learning. The authors highlight five key strategies from learning sciences that instructors often want to use more consistently, but find difficult due to the time needed for material development and implementation. They suggest prompt formulations for different AI models to generate helpful outputs that can assist with incorporating each technique into the classroom.

What did the researchers do and find?

Through a literature review, the authors established why each strategy has value from a cognitive perspective. They then provided examples of prompts instructors could use with AI systems like large language models to quickly create instructional materials to enact the strategies. The prompts were tested with current AI chatbots to demonstrate how the systems might respond and serve as a starting point for instructors wishing to experiment with AI in their pedagogy.

The key finding was that AI need not replace human instructors, but rather augment their expertise through rapidly generating supplemental resources instructors could then vet and employ as they see fit. By intentionally implementing evidence-based techniques with AI’s support, educators could act as “force multipliers” and more efficiently help students learn in impactful ways.

How might you use this research in your teaching?

Key Strategies:

1. Providing numerous examples – Studies show exposure to varied, concrete examples helps students grasp abstract concepts and promote transfer of learning. But creating many tailored examples is time-consuming for instructors. Prompting the AI to generate numerous examples adapted to the student level allows quick access to this useful material.
2. Explaining concepts in multiple ways – Research indicates explanations utilizing analogies, step-by-step breakdowns, and examples aid student understanding. However, tailoring explanations to different student levels requires considerable effort. The AI can efficiently produce varied explanations and analogies to supplement instructor teachings.
3. Low-stakes testing – Short, frequent quizzes and questions provide retrieval practice shown to boost retention and metacognition. But regularly developing quality diagnostic questions aligned to learning goals is difficult. The AI can swiftly generate personalized quizzes incorporating common student misconceptions.
4. Assessing knowledge gaps – Techniques like muddiest point reflections reveal student difficulties but require time-intensive analysis of responses. Prompting the AI to review responses lets instructors quickly gain insights into class confusions and misconceptions.
5. Distributed practice – Interleaving topics and incorporating spaced repetition have benefits for learning, but few course materials are designed this way. The AI helps efficiently create connections between concepts and repetitions spaced over time.

For university teachers and instructors, this research suggests AI may open up new possibilities for operationalizing sound teaching methods linked to deeper learning. When used judiciously and under an educator’s guidance, AI tools show promise not just as timesavers but as a means of giving more students access to pedagogies shown to boost retention and transfer. Moving forward, collaborative work between educational researchers and AI designers could yield even more practical applications of emerging technologies.