AI-powered question detection tools like Question Bot (QBot) show promising potential for enhancing teacher reflection and professional development, with student-teachers reporting it helps them notice patterns in their classroom questioning and make decisions to improve their teaching practices.
Objective: The main goal of this study was to investigate student-teachers' insights into an AI-powered automatic question detection tool (Question Bot or QBot) designed for reflective practice, and to explore their perspectives on potential future AI tools for classroom interaction analysis.
Methods: The research involved six female student-teachers in their final year of a secondary school teacher education program in Sweden who volunteered to use QBot to analyze their own classroom discourse from prerecorded micro-teaching lessons. Semi-structured interviews averaging 57 minutes were conducted after students completed their course and received their grades. The interviews explored the student-teachers' practical use of QBot for identifying and reflecting on questions they asked, their insights into AI partnership in teacher education, and their vision for future AI tools. Data was analyzed using thematic analysis with NVIVO software, generating 86 initial codes that were merged into themes and sub-themes.
Key Findings:
- QBot promotes noticing and pattern identification: All six student-teachers reported that the tool enabled them to notice aspects of their questioning practices and identify patterns, such as overuse of yes/no questions or lack of fact-checking questions.
- QBot facilitates reflection on action: Five of six participants reflected on their actions and teaching based on what they noticed through QBot, considering alternative approaches like asking more genuine questions.
- QBot facilitates reflection for action: Four participants made future-oriented decisions about their teaching based on QBot-enhanced reflections, such as committing to use more fact-checking questions or fewer yes/no questions.
- AI partnership preference: All participants preferred AI-human partnership rather than AI-only feedback, valuing the expertise of human mentors and the dialogue that comes with human interaction.
- Future AI tool suggestions: Student-teachers envisioned complementary AI features that could detect gestures, gaze, silence, voice pitch, feedback patterns, and code-switching to enhance reflection.
Implications: The findings suggest that AI-powered tools like QBot can be effectively integrated into teacher education programs to enhance reflective practice, especially when combined with video-based classroom observations. The study supports the value of AI-human collaboration rather than AI-only approaches, preserving human agency and creativity while leveraging the benefits of automation. It highlights how AI tools can facilitate deeper levels of teacher noticing and reflection, potentially leading to improved classroom discourse and teaching practices.
Limitations: The researcher acknowledges several limitations, including that participants had taken a classroom interaction course before using QBot, which might have amplified its impact. The small sample size (six student-teachers) from the same subject area limits generalizability. Additionally, the interview method with some stimulated recall might have provided certain directions to the student-teachers, and an unguided approach might yield different results.
Future Directions: The author suggests several areas for future research:
- Testing QBot with an unguided approach
- Research with student-teachers at different levels in their education
- Investigating QBot use with more experienced teachers who might bring more practical expertise
- Studies with more diverse participants from different subject areas
- Research that investigates conversations of student-teachers rather than just interviews
- Longitudinal studies to track development over time
- Combining verbal and linguistic aspects of interaction with visual and non-verbal phenomena in AI tools
- Exploring whether transcriptions are necessary for automatic classroom discourse analysis
- Developing better visualization tools for AI feedback
Title and Authors: "Partnering with AI in teacher education? Using an automatic question detection tool to reflect on classroom interaction" by Olcay Sert.
Published On: May 19, 2025
Published By: Journal of Research on Technology in Education
The study positions itself within a growing body of research on AI applications in classroom discourse analysis, noting that while many studies have developed AI models for analyzing classroom interaction, few have applied these models in practice or investigated teachers' actual use of such tools. The researcher emphasizes that rather than using AI technology for teacher assessment and evaluation, which could be problematic, research should focus on creating opportunities for teachers to learn from their own practice through AI assistance.
The author argues that collaboration between human experts and AI tools is crucial for developing rich reflective practices, as it is through dialogic reflections that teachers learn better. AI tools should facilitate such dialogues rather than being the sole provider of feedback. This requires a good level of critical AI literacy for student-teachers, experienced teachers, and teacher educators.
The Question Bot (QBot) tool studied is embedded in the Video Enhanced Observation (VEO) tool and automatically identifies questions asked during lessons, allowing teachers to navigate and reflect on these questions at their own pace. This research represents an important step toward understanding how AI can support teacher development while maintaining the essential human elements of reflective practice.