Reimagining Textbook Learning: An Interactive AI Tutor Approach Using Retrieval-Augmented Generation

ACL ARR 2026 January Submission6545 Authors

Keywords: Retrieval-Augmented Generation, Intelligent Tutoring Systems, Educational NLP, Human-centered AI, Open Educational Resources

Abstract:

We present Textbook Tutor, an interactive AI learning system that transforms static textbook chapters into structured learning experiences through a reproducible Textbook-to-Interaction pipeline. The system converts open educational resources into multiple instructional modes, including storytelling, business case simulations, and interactive challenges grounded in retrieval-augmented generation.
Through controlled pipeline-level evaluations and a mixed-method user study (N=62), we find that instructional structure contributes more to short-term learning gains than increasing model capacity alone, and that lightweight lexical retrieval is sufficient for reliable chapter-level grounding.
Guided by usability findings from an early prototype, we transitioned the system to a web-based deployment supporting both curated single-textbook tutoring and flexible multi-book workflows. Together, our results demonstrate that effective and accessible interactive textbook learning can be achieved without heavy infrastructure.

Paper Type: Long

Research Area: Human-AI Interaction/Cooperation and Human-Centric NLP

Research Area Keywords: human-centered NLP, intelligent tutoring systems, educational applications, retrieval-augmented generation

Contribution Types: NLP engineering experiment, Approaches to low-resource settings, Publicly available software and/or pre-trained models, Data resources, Data analysis

Languages Studied: English

Submission Number: 6545