Using GenAI To Support Retrieval Practice
Generative Artificial Intelligence can generate practice test questions, assess student answers, and provide corrective feedback to enhance student learning and retention.
Presented by:
Jonathan G. Tullis, University of Arizona
Hear it from the author:
Transcript:
Alright, welcome. We were interested in whether Generative Artificial intelligence can support student learning by creating practice test questions and providing feedback. Because we know that taking practice tests is one of the most effective ways to learn and remember information.
Students studied four passages related to their classes. Then, texts were either assigned to a reread condition, or a practice questions from generative AI, where we copied and pasted the readings into ChatGPT and ChatGPT asked students questions about those readings and provided feedback.
Students then took a final test about how much they've learned, either immediately or after 2 days.
And what we found was that students believed they had learned more from rereading than answering practice test questions from generative AI. But they performed better on those final tests when they had taken practice test questions through generative AI.
Showing the generative AI can be a useful tool to support long-term learning of this information.
Thanks.
Key Words:
Generative Artificial Intelligence, Cognition, Retrieval Practice
Abstract:
Retrieval practice enhances learning relative to rereading, but implementing retrieval practice can be challenging. Here, we examine whether Generative Artificial Intelligence (genAI) can generate effective practice questions during study. Texts were assigned to either rereading or AI-supported retrieval practice conditions. In the retrieval practice condition, students answered AI-generated questions about the readings and received immediate feedback. We assessed learning with immediate and delayed multiple-choice tests. On the delayed test, learners in the AI-supported retrieval practice condition answered more questions correctly than students who reread the material. GenAI can effectively instantiate retrieval practice and may offer a scalable tool for promoting durable learning in educational contexts.
Outcomes:
1. Describe one way of using genAI effectively to support learning.
2. Adapt genAI to their own courses.
3. Connect basic memory literature with classroom applications.
References:
Karpicke, J. D., & Blunt, J. R. (2011). Retrieval practice produces more learning than elaborative studying with concept mapping. Science, 331,772–775. doi:10.1126/science.1199327
Karpicke, J. D., & Roediger, H. L., III. (2008). The critical importance of retrieval for learning. Science, 319, 966 –968. doi:10.1126/science.1152408
Rawson, K. A., & Dunlosky, J. (2012). When is practice testing most effective for improving the durability and efficiency of student learning? Educational Psychology Review, 24, 419 – 435. doi:10.1007/s10648-012-9203-1