AI Meets Austen: Towards Human-Robot Discussions of Literary Metaphor
Project Overview
The document explores the application of generative AI in education, specifically through the creation of an embodied conversational robot aimed at enhancing discussions around literary metaphors in fiction. It emphasizes the robot's ability to foster lifelong learning and cognitive engagement, particularly in informal educational contexts. Usability evaluations reveal that participants found the robot to be engaging and likable during multiple interactions, indicating a favorable reception towards the integration of AI companions in educational settings. These findings suggest that generative AI can effectively support learning by facilitating meaningful conversations and stimulating interest in literary analysis, thereby contributing to the broader goals of promoting continuous education and enhancing student engagement.
Key Applications
Embodied conversational robot designed for book discussions
Context: Informal lifelong learning setting, targeting adult readers discussing literature, specifically Jane Austen's 'Pride and Prejudice'.
Implementation: The robot, named Grace, was implemented using a dialogue system that generates questions about metaphors in the text. The system functioned autonomously with a speech recognition technique.
Outcomes: Participants rated the robot positively in terms of engagement and likeability, suggesting it can effectively promote cognitive exercise and discussions about literary content.
Challenges: The system's conversational capabilities are limited to predetermined structures, which may restrict open-ended discussions.
Implementation Barriers
Technical barrier
Limited ability for open-ended conversation, relying on predefined questions which can restrict meaningful engagement.
Proposed Solutions: Future work will focus on enhancing personalization and improving conversation quality to facilitate more natural interactions.
Project Team
Natalie Parde
Researcher
Rodney D. Nielsen
Researcher
Contact Information
For information about the paper, please contact the authors.
Authors: Natalie Parde, Rodney D. Nielsen
Source Publication: View Original PaperLink opens in a new window
Project Contact: Dr. Jianhua Yang
LLM Model Version: gpt-4o-mini-2024-07-18
Analysis Provider: Openai