View in EDS PDF Full Text

From ChatGPT to classroom learning: Exploring the role of teacher mediation in AI-supported education.

Saved in:
Bibliographic Details
Title: From ChatGPT to classroom learning: Exploring the role of teacher mediation in AI-supported education.
Authors: Borges, Inês1 iborges@iscac.pt, Caridade, Cristina M. R.2, Sebastião, Cláudia1, Pereira, Verónica3
Source: Australasian Journal of Educational Technology. 2026, Vol. 42 Issue 2, p116-137. 22p.
Subject Terms: *Mathematics students, *Lesson planning, *Academic support programs, *Intelligent tutoring systems, *Active learning, ChatGPT
Abstract: This study investigated how undergraduate students in a first-year applied mathematics course engaged with artificial intelligence (AI) tools and how teacher mediation influenced learning outcomes. Adopting a lesson study approach, grounded in the framework of instrumental orchestration, the research followed 127 students as they tackled a new integration method. Results from the autonomous phase revealed a verification deficit, despite high AI usage habits, only 15.6% of groups produced correct or mostly correct solutions using tools like ChatGPT and only 14.2% of students reported feeling confident. Following a structured intervention focused on equipping students with foundational mathematical knowledge and verification strategies, Wilcoxon signed-rank tests indicated that group performance rose significantly to 74.2% (Z = 4.62, p < .001, r = .82) and individual confidence increased to 36.3%. Furthermore, a final problem posing phase showed that shifting students from consumers to evaluators enabled groups to construct problems through prompt refinement and validation. These findings suggest that while AI tools foster autonomy, they require explicit teacher scaffolding to transform blind trust into critical understanding. The study highlights implications for course design, teacher training and institutional strategies for integrating AI into higher education, demonstrating the value of lesson study for examining how emerging technologies reshape pedagogical practice. Implications for practice or policy: • First-year students require foundational structural knowledge before autonomous AI engagement to mitigate the verification deficit. • Teacher mediation must evolve from simple content exposition to epistemic scaffolding, explicitly equipping students with the structural criteria required to audit and validate algorithmic outputs. • Course designs should integrate problem posing tasks. Shifting students from consumers to evaluators compels them to create and refine prompts, validating structural understanding and fostering the internalisation of concepts. [ABSTRACT FROM AUTHOR]
Copyright of Australasian Journal of Educational Technology is the property of Australasian Journal of Educational Technology (AJET) and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
Database: Education Research Complete
Description
Abstract:This study investigated how undergraduate students in a first-year applied mathematics course engaged with artificial intelligence (AI) tools and how teacher mediation influenced learning outcomes. Adopting a lesson study approach, grounded in the framework of instrumental orchestration, the research followed 127 students as they tackled a new integration method. Results from the autonomous phase revealed a verification deficit, despite high AI usage habits, only 15.6% of groups produced correct or mostly correct solutions using tools like ChatGPT and only 14.2% of students reported feeling confident. Following a structured intervention focused on equipping students with foundational mathematical knowledge and verification strategies, Wilcoxon signed-rank tests indicated that group performance rose significantly to 74.2% (Z = 4.62, p < .001, r = .82) and individual confidence increased to 36.3%. Furthermore, a final problem posing phase showed that shifting students from consumers to evaluators enabled groups to construct problems through prompt refinement and validation. These findings suggest that while AI tools foster autonomy, they require explicit teacher scaffolding to transform blind trust into critical understanding. The study highlights implications for course design, teacher training and institutional strategies for integrating AI into higher education, demonstrating the value of lesson study for examining how emerging technologies reshape pedagogical practice. Implications for practice or policy: • First-year students require foundational structural knowledge before autonomous AI engagement to mitigate the verification deficit. • Teacher mediation must evolve from simple content exposition to epistemic scaffolding, explicitly equipping students with the structural criteria required to audit and validate algorithmic outputs. • Course designs should integrate problem posing tasks. Shifting students from consumers to evaluators compels them to create and refine prompts, validating structural understanding and fostering the internalisation of concepts. [ABSTRACT FROM AUTHOR]
ISSN:14493098
DOI:10.14742/ajet.11056