How to Improve the Explanatory Power of an Intelligent Textbook: a Case Study in Legal Writing

Area of Science:

• Artificial Intelligence in Education
• Natural Language Processing
• Cognitive Science

Background:

• Explaining complex information intuitively is a significant challenge in education.
• Current educational materials often lack user-centered, adaptive explanations.
• Existing AI tools struggle to leverage vast content libraries effectively for personalized learning.

Purpose of the Study:

• To investigate how explanatory AI can improve the exploitation of rich content libraries.
• To develop an interactive and adaptive e-book system (YAI4Edu) based on question-answering technology.
• To evaluate the effectiveness of YAI4Edu-enhanced textbooks compared to traditional and general-purpose explanatory tools.

Main Methods:

• Developed YAI4Edu software using question-answering technology and knowledge graph extraction.
• Framed explanation as illocutionary question-answering, based on Ordinary Language Philosophy.
• Conducted a within-subjects user study with over 100 students using legal memorandum writing excerpts.

Main Results:

• YAI4Edu generates on-demand, expandable explanations tailored to user needs.
• Students rated YAI4Edu's explanations significantly higher than baseline alternatives (P < .005).
• The AI-driven system proved more effective than random and general-purpose explanatory tools.

Conclusions:

• YAI4Edu offers a novel approach to interactive and adaptive learning.
• AI-powered educational tools can significantly improve the pedagogical value of digital content.
• This technology has the potential to revolutionize how students engage with and understand complex subject matter.