AI language polishing alone can flip human-authored scholarly texts from human-classified to AI-classified¶
Claim · All claims
Evidence · 1 study · reviewr3·q3
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.
q3 i? After ChatGPT-4o was instructed to enhance readability and flow while preserving meaning, 75–85% of previously human-classified research letters were classified as AI-generated, including 15–25% at high confidence. → Victor Angelier 2026
Evidence¶
Victor Angelier 2026¶
Victor Angelier. (2026). The pitfalls of AI detection in academic writing: bias, false positives, and the need for inclusive assessment. AI and Ethics. https://doi.org/10.1007/s43681-026-01376-w
q3 · i? · review · r3
The review reports Wang et al.'s experiment on JAAD International research letters, where an intervention designed to preserve "original meaning and intent" flipped classifications from human to AI-generated. The authors reported residual meaning changes in two to three sentences per letter among ten highly flagged letters.
"Before polishing, 97–100% of the human-authored originals were classified as human; after polishing, 75–85% of the versions were classified as AI-generated, including 15–25% at high confidence."
Discussion¶
Related Claims¶
- A detector can achieve high aggregate accuracy while producing materially different over-detection rates across author groups — 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
- In a documented exploratory audit, one human-authored manuscript received materially different classifications from five commercial detectors, spanning 0% human to Human Generated — related
- AI-generated rewrites of cover letters were rated higher for writing quality and more likely to secure hypothetical interviews than rewrites by professional human editors — related
- Editors observe GenAI has improved surface correctness of manuscripts and increased submissions, but this does not lead to better research quality — related
- All four detection tools classified 100% of fully human-written papers as true negatives, with false positives nearly absent — reports the opposite
- AI-generated-text detectors are reported as ineffective against current sophisticated language models, and students can alter text to evade detection — related
- In a paired exploratory comparison, extensive machine rewriting by a commercial humanizer materially degraded the manuscript while changing the detection score only slightly — related
Page metadata
| Field | Value |
|---|---|
type |
claim |
id |
polishing-flips-detector-classification |
title |
AI language polishing alone can flip human-authored scholarly texts from human-classified to AI-classified |
description |
AI language polishing alone can flip human-authored scholarly texts from human-classified to AI-classified |
status |
review |
evidence_strength |
moderate |
newest_evidence |
2026 |
recheck_by |
2029 |
generated |
by process:wiki-ingest on 2026-10-10 |
trust |
unverified — no one has confirmed this page's content |
sources |
1 — victor-angelier-2026 |
Frontmatter is the page's machine-readable record — see the Schema & Guide.