Skip to content

AI book discussion companion: a fixed peer-role conversational agent using Whisper speech-to-text and GPT-3.5

Design · All designs
Evidence · 2 claims (2 for) · 1 study (1 causal), q2 · 0 of 1 report an effect size · 2 claims rest on one study

Description

A conversational AI system developed by the research team for elementary book talk. "This system utilized Ope - nAI's Whisper API for accurate speech-to-text transcription of student utterances and the GPT-3.5 API for natural language understanding and the generation of contextu - ally relevant responses, delivered via both text and synthesized speech (Web Speech API)." For the comparative study it was configured to act as a peer-like, non-evaluative partner following the 4F framework sequentially, and it operated in this fixed role throughout AI sessions. The system automatically logged session durations and generated complete transcripts used for analysis.

Design Implications

Context

Requirements

  • Speech-to-text and language-model APIs for transcription and response generation
  • A fixed peer-role system prompt following the 4F framework sequentially

Constraints

  • Operated in a single fixed peer role for the study, without context-aware role adaptation

Target Learners

  • Elementary school students in Grades 4 and 5

Learning Goals

  • Book talk dialogue about self-chosen books
  • Reflection prompted in a non-evaluative manner

Claims

Examples

-

Key Sources

  • Liao, C.-Y. (2026). Beyond a single role: Justifying a role-adaptive framework for ai companions through a comparative study in elementary book talk. Education and Information Technologies. https://doi.org/10.1007/s10639-026-13958-6
Page metadata
Field Value
type design
id ai-book-discussion-companion-whisper-gpt35
title AI book discussion companion: a fixed peer-role conversational agent using Whisper speech-to-text and GPT-3.5
description A conversational AI system developed by the research team for elementary book talk.
status review
generated by process:settle-candidates on 2026-10-11
trust unverified — no one has confirmed this page's content
sources 1 — liao-2026

Frontmatter is the page's machine-readable record — see the Schema & Guide.