CoChat

Mentorship in the Age of Generative AI: ChatGPT to Support Self-Regulated Learning of Pre-Service Teachers Before and During Placements

Open access
Nguyen (Ruby), Ngoc Nhu, Barbieri, Walter · 2025 · Education Sciences

The DOI matches its Crossref record; 76 of the 78 references it cites were found in a registry or an index; 19 of the 39 sources its claims cite were checked for support.

Abstract

This study investigates the integration of mentorship, self-regulated learning (SRL), and generative artificial intelligence (gen-AI) to support pre-service teachers (PSTs) before and during work-integrated learning (WIL) placements. Utilising the Mentoring and SRL Pyramid Model (MSPM), it examines how mentors’ dual roles as coaches and assessors influence PSTs’ SRL and explores to what extent gen-AI can assist PSTs in meeting the demands of WIL placements. Quantitative and qualitative data from 151 PSTs, including survey, interview, placement scores, and mentor feedback were analysed using statistical correlation analysis and thematic analysis to reveal varied mentorship approaches. Gen-AI tools are highlighted as valuable in enhancing PSTs’ SRL, providing tactical and emotional guidance where traditional mentorship is limited. However, challenges remain in gen-AI’s ability to navigate complex interpersonal dynamics. The study advocates for balanced mentorship training that integrates technical and emotional support, and equitable access to gen-AI tools. These insights are critical for educational institutions aiming to optimise PST experiences and outcomes in WIL through strategic integration of gen-AI and mentorship.

Overview

CoChat Summary

TL;DR

This mixed-methods study uses the Mentoring and Self-Regulated Learning Pyramid Model (MSPM) to examine how generative AI (ChatGPT) supported 151 pre-service teachers' self-regulated learning before and during five-week work-integrated learning placements. Gen-AI functioned as a supplementary scaffold for reflection, feedback gaps, and instructional planning, but could not replace human mentors for psychosocial support, behaviour management, or stress reduction.

Key findings

  • Of 151 PSTs, 139 (92%) passed placement. Scores were highest in APST 2.6 (ICT integration, Mean = 4.3), APST 3.1 (lesson planning, Mean = 4.1) and APST 3.5 (classroom communication, Mean = 4.05), and lowest in APST 4.3 (managing challenging behaviour) and APST 7.4 (professional networks), each Mean = 2.6.
  • PSTs rated gen-AI highly for suggesting instructional adjustments (Q17, Mean = 4.27, SD = 0.76), filling feedback gaps when mentors were unavailable (Q14, Mean = 4.25, SD = 0.73), reflection (Q16, Mean = 4.11, SD = 0.79), goal setting (Q18, Mean = 4.19, SD = 0.81) and emotional support (Q13, Mean = 4.09, SD = 0.79). Gen-AI engagement correlated moderately with SRL confidence (r = 0.41, p < 0.01).
  • Gen-AI was rated poorly for clarifying mentor expectations (Q12, Mean = 3.21, SD = 0.88), improving behaviour management (Q15, Mean = 3.00, SD = 0.85) and reducing placement stress (Q20, Mean = 2.41, SD = 0.92). Interview and mentor feedback flagged risks of over-reliance and passive decision-making.
  • Co-regulated mentor–PST AI use emerged in some cases (e.g., MT4 adopting AI for lesson structuring after a PST demonstrated it), but such instances were "limited and context-dependent."

Methods

  • Mixed-methods triangulated design: convenience sample of 151 PSTs in Initial Teacher Education at the University of Adelaide, all completing a six-week preparatory course (including AI literacy training with ChatGPT 3.5/4o/4o mini) followed by a five-week secondary school placement; data collected over six months.
  • Data sources: a 20-item 5-point Likert survey aligned to the MSPM (mentor support, SRL strategies, gen-AI usage), 1:1 semi-structured interviews with 12 purposively sampled PSTs, and placement reports scored by mentor teachers against the 37 Australian Professional Standards for Teachers (APST) on a 5-point scale.
  • Analysis: statistical correlation analysis plus reflexive thematic analysis (Braun & Clarke, 2022) of interview transcripts and deidentified mentor qualitative feedback.

Limitations

  • Stated by the paper: gen-AI's effectiveness was context-dependent, with uneven access to AI tools across WIL settings, an absence of clear institutional guidelines for AI use in schools, and limited instances of AI-facilitated co-regulation; the authors note placement scores alone "cannot isolate the effect of gen-AI."
  • Evident in the text: no control or comparison group, so causal claims about gen-AI and SRL/placement outcomes are not supported; convenience sampling limits generalisability; and reported statistics are internally inconsistent (Q16 appears as Mean = 4.11, SD = 0.79 in one section and Mean = 4.22, SD = 0.71 in another; Q20 appears as Mean = 2.41 in the results text but 3.41 in Appendix A).

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.3390/educsci15060642

Cites

76 of 76 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Verification

Claims the paper makes

15 Valid · 4 Not valid · 10 Not checked · 158 Own results and interpretations

96 more claims · 6 with their cited sources checked

Create a free account to see all 187 claims and every source verdict.

See all claims — free account
  • It may state two separate claims as one. 48
  • Probably not a substantial claim. 46
  • The quote only partly states this claim. 1

Create a free account to see these claims.

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

4 of 4

Citations

Cites

Show all 76 references

Research works printed but not linked: 4

  1. Edwards, 2015, Praxis shock: A real challenge for novice teachers, Journal of Teacher Education, 66, 213
  2. 2021, La relation mentorale, Journal of Humanistic and Social Studies, 12, 109
  3. Nam, C. S., Jung, J.-Y., and Lee, S. (2022). Human-centered artificial intelligence: Research and applications, Elsevier Science & Technology. [1st ed.].
  4. Nigate, 2023, Mentoring during school practicum: Mentor-mentee relationship, roles assumed, and focus of feedback, International Journal of Work-Integrated Learning, 24, 491

Cited by

Metadata

Authors
Nguyen (Ruby), N. N., & Barbieri, W.
Affiliations
  • Office of the Deputy Vice-Chancellor (Education), The University of Sydney, Sydney, NSW 2006, Australia
  • The University of Adelaide (AU)
Venue
Education Sciences
Publisher
MDPI AG
Published
2025-05-23
Volume
15
Issue
6
Pages
642
ISSN
2227-7102 (ISSN-L)
OpenAlex
W4410622615
Cited by
15
References
78
Citations by year
2026: 9 · 2025: 6
FWCI
46.88
Type
journal article
Language
en
Open access
Locations
  • Education Sciences (publishedVersion, cc-by)
  • Adelaide Research & Scholarship (AR&S) (University of Adelaide) (submittedVersion, cc-by)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
Topics
  • Social Sciences › Education › Teacher Education and Leadership Studies
  • Social Sciences › Education › Artificial Intelligence in Education
  • Social Sciences › Education › Mentoring and Academic Development
Keywords
mentoring, student self-regulation, generative AI, pre-service teachers, work-integrated learning, teacher placement, ChatGPT, mentor training, emotional support

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

The claim verdicts this report counts are experimental: their accuracy is not yet measured.

A language model drew the claims from the paper on 2026-10-05, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-10-05 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Cited sources are checked with a language model’s help (gpt-6-sol).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.