In a single-institution study, the joint system of Turnitin detection plus faculty judgement formally reported only 54.5 per cent of AI-generated submissions¶
Claim · All claims
Evidence · 1 study · theoreticalr3·q2
Currency · newest study 2026 · not yet checked against its sources · not reviewed by a person · recheck by 2029: tool and model accuracy changes quickly
Subclaims¶
Reading the evidence codes
q is the design tier, i the impact, n the sample, and the kind with its r
the rigour judged within that kind: what each code means.
This claim: no standardized effect is recorded in observations/ for it yet, so no benchmark reading is possible.
q2 i? Perkins et al. (2024) report that although Turnitin's AI detection flagged 91 per cent of experimental GPT-4 submissions, the tool identified only 54.8 per cent of the AI-generated content within them, and faculty formally reported only 54.5 per cent of submissions through the misconduct process. → Craig Wright 2026
Evidence¶
Craig Wright 2026¶
Craig Wright. (2026). Transcription is not generation: distinguishing non-generative AI tool use from academic misconduct in higher education assessment. International Journal for Educational Integrity. https://doi.org/10.1007/s40979-026-00234-w
q2 · i? · theoretical · r3
The paper reports, citing Perkins et al. (2024), an evaluation at a UK-curriculum business school of twenty-two experimental GPT-4 submissions marked by fifteen academic staff alongside 963 genuine student submissions. The tool 'identified only 54.8 per cent of the AI-generated content within those flagged submissions'.
"the tool identified only 54.8 per cent of the AI-gener -ated content within those flagged submissions, despite the entirety of each experimental submission having been AI-generated; faculty, even with the Turnitin AI score visible in the marking interface, formally reported only 54.5 per cent of the experimental submis -sions through the academic-misconduct process."
Discussion¶
Related Claims¶
- On fully AI-generated papers produced with ChatGPT Deep Research, Pangram achieved 65% strict accuracy while Turnitin, GPTZero and Copyleaks classified essentially all papers as false negatives — related
- Universities disabling AI detection cite false-positive risk at scale, and detector influence persists in individual grading, journal screening, and institutions without policy revision — related
- ChatGPT produced submissions judged of passing quality for 36 of 40 undergraduate psychology assessments (90%) — related
- Applied to 1,163 real master's theses, Pangram flagged 529 (45.5%) as containing AI-generated content, mostly at low-to-moderate levels — related
- Universities converge on caution toward AI-detection systems, warning against using detection scores as dispositive evidence of misconduct — related
- A linear mixed-effects model of detection error showed significant paper type, tool, and interaction effects, with underestimation largest for fully AI-generated papers and Pangram consistently producing the smallest errors — related
- Popular plagiarism detection tools detect plagiarism in AI-assisted essays in less than 15% of cases — related
Page metadata
| Field | Value |
|---|---|
type |
claim |
id |
turnitin-joint-system-54-5-percent-reporting |
title |
In a single-institution study, the joint system of Turnitin detection plus faculty judgement formally reported only 54.5 per cent of AI-generated submissions |
description |
In a single-institution study, the joint system of Turnitin detection plus faculty judgement formally reported only 54.5 per cent of AI-generated submissions |
status |
review |
evidence_strength |
weak |
newest_evidence |
2026 |
recheck_by |
2029 |
generated |
by process:wiki-ingest on 2026-10-12 |
trust |
unverified — no one has confirmed this page's content |
sources |
1 — craig-wright-2026 |
Frontmatter is the page's machine-readable record — see the Schema & Guide.