Admissions officers can often discriminate AI-written from human-written essays, achieving an AUC of 0.70, well above chance but below commercial detectors¶
Claim · All claims
Evidence · 1 study · causalr2·q2
Currency · newest study 2026 · not yet checked against its sources · not reviewed by a person
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? Human judgments were directionally informative but imprecise: among essays rated on the human side of the scale, 70.4% were classified as human by GPTZero, and among those rated on the AI side, 67.8% were classified as AI-generated. → Calvin Isley 2026
Evidence¶
Calvin Isley 2026¶
Calvin Isley, Johann D. Gaebler, and Sharad Goel. (2026). AI-written admissions essays are widespread but penalized. arXiv preprint. https://arxiv.org/abs/2609.22549 doi:10.48550/arxiv.2609.22549
q2 · i? · causal · r2
Analysis relating officers' perceived AI likelihood to GPTZero provenance classifications (Fig. 4). Judgments were "directionally informative but imprecise": 70.4% of human-side essays were human-classified and 67.8% of AI-side essays were AI-classified.
"For example, among essays rated on the human side of the scale (1–3), 70.4% were classified as human-generated by GPTZero, and among those rated on the AI side, 67.8% were classified as AI-generated by GPTZero."
Discussion¶
Related Claims¶
- Independent evaluation found commercial AI-text detectors neither sufficiently accurate nor sufficiently reliable for high-stakes use — related
- Admissions staff rate essays they perceive as AI-generated lower than essays they believe are human-written, even after adjusting for measured essay quality and actual AI authorship — related
- By the 2025 admissions cycle, a majority of applicants to the studied public policy master's program submitted at least one primarily AI-generated essay despite an explicit prohibition — 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
- In a human Turing test, participants could not reliably distinguish BEAGLE traces from real student data (52.8% accuracy; d′ = 0.15 within the ±0.3 equivalence bound, pTOST = 0.038) — related
- AI access raises unaided essay quality in both sessions, while AI-generated text detectable in Session One essays disappears by Session Two — related
- Comparative studies report that AI detection tools do not reliably distinguish human-written from AI-generated text — related
- Participants fail to distinguish undisclosed AI teammates from human teammates above chance in synchronous text-based group interaction — related
Page metadata
| Field | Value |
|---|---|
type |
claim |
id |
admissions-officers-detect-ai-essays-auc-070 |
title |
Admissions officers can often discriminate AI-written from human-written essays, achieving an AUC of 0.70, well above chance but below commercial detectors |
description |
Admissions officers can often discriminate AI-written from human-written essays, achieving an AUC of 0.70, well above chance but below commercial detectors |
status |
review |
evidence_strength |
moderate |
newest_evidence |
2026 |
generated |
by process:wiki-ingest on 2026-10-10 |
trust |
unverified — no one has confirmed this page's content |
sources |
1 — calvin-isley-2026 |
Frontmatter is the page's machine-readable record — see the Schema & Guide.