CHI 2022 Paper on Human-AI Collaborative Agent for Supporting Interactive Storybook Reading

I’m excited to share that we have another paper accepted to CHI 2022 after R&R, “StoryBuddy: A Human-AI Collaborative Agent for Parent-Child Interactive Storytelling.” This paper is a collaboration with researchers at U of Notre Dame, IBM, RPI, UW, and GT. We appreciate the very constructive feedback from reviewers that has helped improve the paper.

This paper presents the design and evaluation of StoryBuddy, a system that utilizes automatic question generation to support parents/teachers to engage children in story-related dialogue. The automatic question generation was enabled by a benchmark educational question-answer dataset, “FairytaleQA”, my team has created.

Unlike many fully automatic AI systems, StoryBuddy is designed to keep “parents/teachers in the loop” so that they can be actively involved in the AI-aided content generation processes (through selecting stories, configuring question types, etc.). We believe that keeping parents/teachers in the loop is important to ensure that the reading content generated by StoryBuddy meets the needs of individual children.

I look forward to sharing this paper at CHI2022 in April! I will also be presenting my other paper “Elinor’s Talking to Me!”: Integrating Conversational AI into Children’s Narrative Science Programming.