Does an AI-Supported Teacher Professional Learning System Improve Student Performance?

• In Event: AI Across Educational Contexts

Abstract

Self-paced online professional development courses for k-12 educators have gained significant attention owing to their convenience and flexibility regarding time and location. However, an issue often arises within these self-paced online learning systems is the lack of frequent, timely interaction and feedback for educators. This study addresses this issue by developing a computerized program for middle school mathematics teachers. The program offers personalized, real-time interactions and feedback through a virtual agent, utilizing natural language processing. A randomized experiment demonstrated that this program significantly improved students’ mathematics performance, even when controlling for their teachers’ baseline knowledge. These findings underscore the potential of integrating an automated, interactive feedback tool backed by artificial intelligence into self-paced online teacher professional development courses.

Authors

• Yasemin Copur-Gencturk, University of Southern California

• Jingxian Li, University of Southern California