AI evidence extraction

Authorship Statement for Generative Artificial Intelligence: Assuring Trust and Accountability

Joseph Crawford, Alison Purvis, Averil Grieve · Journal of University Teaching and Learning Practice · 2026

AI-generated evidence extraction, verified across multiple analytical personas. Not a substitute for the peer-reviewed original.

9/10

Relevance

1/4

Quality (LMQS)

Evidence

Citations

17.49

FWCI

Top 10%

Impact

This is an AI analysis. Read the peer-reviewed original at the publisher: https://doi.org/10.53761/v16abt43

Methodology & findings

Study design

Hermeneutic analysis and argumentative policy development.

Main result

The paper establishes that "authorship transparency is an analogous form of clarity that underpins shared trust in the paper's intellectual and ethical foundations." The authors identify that GenAI has evolved from a discrete writing aid to an embedded and potentially invisible feature of scholarly work, creating ambiguity about acceptable use and accountability. The study finds that "the rapid evolution of GenAI capabilities and GenAI's integration into everyday research and publication workflows has outpaced policy development across the sector," necessitating updated editorial standards with six core commitments covering disclosure, substantive contribution, human judgment, synthetic media, reflexivity, and accountability.

Research paradigm

Interpretive/Argumentative - normative editorial policy development

Author conclusions

The authors conclude that "this position formalises a shift from artefact-based inference to accountability-based transparency in a post-detection scholarly environment." They state that the six commitments "operationalise that objective by making expectations explicit, consistent, and defensible across submission, review, and publication processes." The authors emphasize: "The purpose is not to resist technological change, but to preserve trust in the scholarly record by keeping responsibility anchored to identifiable human authors and by ensuring GenAI use is transparent, bounded, and accountable."

Risk of bias

This is a policy document rather than empirical research. However, potential bias sources include: the authorial perspective of journal editors (institutional position bias), focus on a single journal's standards (limited generalizability), and the normative framing that may reflect specific disciplinary values in higher education rather than universal principles applicable across all research domains.

Limitations

The authors note that "detection-led approaches to integrity" have significant limitations, as "tool outputs are increasingly indistinguishable from human text" and the sector "cannot rely on observable artefact cues as a stable basis for judgement about authorship." Additionally, they acknowledge that "even though tools may initially appear helpful in critiquing other's work..
they do not hold accountability for errors, bias, omissions, or misuse of confidential material." The authors also identify that some researchers may "over-disclose in ways that invite unnecessary scrutiny
b) under-disclose in ways that create reputational and ethical risk
or c) avoid disclosing altogether because the boundaries feel contested."

Open questions raised

The authors identify that "it remains ambiguous to authors as to how to best engage in ethical GenAI use in publishing and writing" and note the absence of "a shared authorship and editor expectation that we can all interpret coherently and consistently."
The paper identifies that there is "a lack of clear practice to guide authors in how to describe GenAI tool use" and notes "ambiguity to authors as to how to best engage in ethical GenAI use in publishing and writing." The authors observe that "The rapid evolution of GenAI capabilities and GenAI's integration into everyday research and publication workflows has outpaced policy development across the sector."
The authors identify that there is currently "a lack of clear practice to guide authors in how to describe GenAI tool use" and note the absence of "a shared authorship and editor expectation that we can all interpret coherently and consistently." They also identify the gap that "the scholarly landscape that now prevails might be seen as existing on a spectrum between those fearful or avoidant of GenAI use to those who use it without declaration or concerns in relation to the ethicality of use."

Extracted from: pdfAgreement 81%

Explore related topics

Responsible AI Use Research Integrity Position Paper