Bidirectional Feedback-Based Personalization of Learning using Multi-tier AI: A Real-World Assessment of its Efficacy in Classrooms

Authors

  • Saptarshi Basu Georgia Institute of Technology
  • Jayson Brown Georgia Institute of Technology
  • Cherie Lum Georgia Institute of Technology
  • Junsoo Park Georgia Institute of Technology
  • Ashok K. Goel Georgia Institute of Technology

DOI:

https://doi.org/10.1609/aaaiss.v5i1.35553

Abstract

Jill Watson is an example of an intelligent conversational AI Teaching Assistant that has been deployed across 24 class sections in different institutions, with 1102 unique student participants and over 17000 questions from Fall 2023 to present. Jill Watson’s RAG-based architecture built around OpenAI ChatGPT and user study results addresses some of the concerns related to domain knowledge, deployment, and data collection in online classrooms. In this work, a 3-tiered framework for personalization in online education using AI tools to enable a human-AI personalization loop, grounded in real-world human feedback is proposed.
AAAI Summer Symposium 2025 Proceedings Cover

Downloads

Published

2025-05-28

How to Cite

Basu, S., Brown, J., Lum, C., Park, J., & Goel, A. K. (2025). Bidirectional Feedback-Based Personalization of Learning using Multi-tier AI: A Real-World Assessment of its Efficacy in Classrooms. Proceedings of the AAAI Symposium Series, 5(1), 50–51. https://doi.org/10.1609/aaaiss.v5i1.35553

Issue

Vol. 5 No. 1: Proceedings of the 2025 AAAI Spring Symposium Series

Section

Current and Future Varieties of Human-AI Collaboration