CoChat

Co-creating custom GPTs: an autoethnographic study of undergraduate students as partners in generative AI innovation

Open access

The DOI matches its Crossref record; 35 of the 35 references it cites were found in a registry or an index; 14 of the 50 sources its claims cite were checked for support.

Abstract

This study explores how undergraduate students co-created and evaluated custom Generative AI (GenAI) tools within a students-as-partners (SaP) framework. Undergraduate participants, within a Summer Scholarship program and in a placement coursework unit, collaborated with academic staff to design, test, and refine custom Generative Pre-trained Transformer chatbots (cGPTs), aligned with pedagogical goals. Using autoethnographic methodology and thematic analysis of student-authored reflections, we examine how students experienced this process as a form of partnership-driven learning. An inductive analysis of summer scholarship autoethnographic logs reveals insights into the iterative, unpredictable, and creative nature of working with GenAI. Abductive analysis of placement reflections, guided by SaP principles, highlights themes of agency, reciprocity, reflexivity, identity shift, authenticity, adaptability, and technology-pedagogy nexus. Synthesising both perspectives, we argue that co-creation with GenAI can extend the SaP paradigm to include GenAI tools. We also reflect on the ethical and pedagogical implications of GenAI in higher education and advocate for student voice as central to responsible innovation. By embedding partnership pedagogy into GenAI co-creation, this study offers a model for more inclusive, critical, and transformative engagement with emerging technologies in university settings.

Overview

CoChat Summary

TL;DR

This autoethnographic study examined undergraduate students co-creating seven custom GPTs for six pharmaceutical science units within a students-as-partners framework. It found that co-creation supported agency, reciprocity, reflexivity, identity shift, authenticity, adaptability, and technical-pedagogical integration, and proposes extending SaP to triadic student–educator–GenAI partnerships.

Key findings

  • Summer Scholarship SPAR reflections portrayed cGPT development as iterative, unpredictable, and creative: minor wording edits could clearly affect GPT responses or have no apparent effect, and “if a broad idea is included, then it will likely appear in a response.”
  • Placement reflective reports aligned with SaP principles, including agency and ownership, reciprocity and negotiation, identity shift, reflexivity, adaptability arising from authenticity, and technical-pedagogical integration; one student reported “two custom GPTs with 400 collected responses and 800 individual scores.”
  • The project produced seven cGPTs for six undergraduate units, involving five undergraduate students and five academics; development lasted two to six weeks depending on topic, prior experience, and available resources.
  • The authors argue SaP can be extended to include GenAI as a mediating partner in “triadic partnerships,” while cautioning that GenAI lacks agency, intentionality, or moral responsibility.

Methods

  • Qualitative autoethnographic/reflective design at Monash University; five undergraduate students and five academics co-created seven cGPTs for six units, with students recruited through the Summer Scholarship Programme and the BPS3072 Advanced Professional Experience placement.
  • Data sources were Summer Scholarship SPAR autoethnographic logs (Stimulus, Processing, Analysis, Reflection) and placement written reflective reports of approximately 500–1000 words guided by Gibbs’ reflective cycle, supported by contextual artefacts such as prompts, knowledge bases, generated responses, evaluation scores, and meeting discussions.
  • Analysis used reflexive thematic analysis: Summer SPAR logs were analysed inductively by student co-creators, while placement reflections were analysed abductively by academic researchers using a SaP codebook; trustworthiness was supported by member checking, peer debriefing, thick description, and reflexive discussion.

Limitations

  • Stated: the dataset had a small number of student participants, representing a self-selected, high-performing cohort likely predisposed to reflective and independent learning, so findings may not generalise to broader undergraduate populations.
  • Stated: the dual role of students as both participants and reflectors may have shaped reflections toward more constructive or affirming narratives, introducing a degree of positive bias.
  • Stated: the study focused on English-speaking STEM units at a single institution, so institutional culture, disciplinary norms, and language fluency may have influenced engagement; the paper also notes it does not claim to be a full collaborative autoethnography across all participants.

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

Integrity signals

Registry

Crossref · 10.3389/feduc.2026.1818781

Cites

35 of 35 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Figures

Image 1
From the PDF

Verification

Claims the paper makes

13 Valid · 2 Not valid · 22 Not checked · 167 Own results and interpretations

83 more claims

Create a free account to see all 204 claims and to verify the 2 references still open.

See all claims — free account
  • Probably not a substantial claim. 63
  • It may state two separate claims as one. 69
  • The quote only partly states this claim. 2

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.

1 of 1

Citations

Cites

Research works printed but not linked: 7

  1. Chang, 2012, Collaborative Autoethnography
  2. Cook-Sather, 2014, Engaging Students as Partners in Learning and Teaching: A Guide for Faculty
  3. Gibbs, 1988, Learning by Doing: A Guide to Teaching and Learning Methods
  4. Healey, 2014, Engagement Through Partnership: Students as Partners in Learning and Teaching in Higher Education
  5. Paul, 2006, The Thinker’s Guide to the art of Socratic Questioning
  6. Streeting, 2009, Rethinking the Values of Higher Education: Consumpton, Partnership, Community?
  7. Wynants, 2025, Ethical Principles AI Framework for Higher Education

Metadata

Authors
Lo, S., Lea, T., Mavromoustakos, A., Joseph, B., Yii, M., Exintaris, B., Karunaratne, N., Pearson, D. L., Ritchie, T., Yuriev, E., & Pyun, J.
Affiliations
  • Monash University (AU)
  • University of Warwick (GB)
Corresponding author
  • Yuriev, Elizabeth
  • Pyun, Jae
Venue
Frontiers in Education
Published
2026-06-24
Volume
11
ISSN
2504-284X (ISSN-L)
OpenAlex
W7165739316
Cited by
0
References
33
FWCI
0.00
Type
journal article
Language
en
Open access
Locations
  • Frontiers in Education (publishedVersion, cc-by)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion)
Topics
  • Social Sciences › Education › Higher Education Practises and Engagement
  • Social Sciences › Education › Artificial Intelligence in Education
  • Social Sciences › Education › Digital Education and Society
Keywords
generative AI, custom GPTs, students as partners, co-creation, undergraduate students, chatbots, partnership pedagogy, student voice, student agency, responsible innovation, AI ethics in higher education, student reflection

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 2026-10-10, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-10-10 (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.