A human assessor agreed with the majority of GPT-generated feedback comments on assessment aspects that instructors had not commented on¶
Claim · All claims
Evidence · 1 study · designr2·q2
Currency · newest study 2024 · not yet checked against its sources · not reviewed by a person · recheck by 2027: 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? 90.10% of GPT-4 feedback comments on aspects instructors did not cover obtained assessor agreement, versus 73.33% for GPT-3.5. → Dai 2024
Evidence¶
Dai 2024¶
Wei Dai, Yi-Shan Tsai, Jionghao Lin, Ahmad Aldino, Hua Jin, Tongguang Li, Dragan Gašević, Guanliang Chen. (2024). Assessing the Proficiency of Large Language Models in Automatic Feedback Generation: An Evaluation Study. Computers and Education: Artificial Intelligence, 7, 100299. Preprint submitted to Elsevier. https://doi.org/10.1016/j.caeai.2024.100299
q2 · i? · design · r2
Supplementary validation in which a human assessor rated GPT comments on aspects the original instructor had not addressed (Table 9), choosing agree, partially agree, or disagree. "90.10% of these pieces of feedback obtained agreement from the human assessor" for GPT-4; 73.33% for GPT-3.5.
"more feedback pieces generated by GPT-4 (101 out of 103 pieces of feedback) contained comments that were not commented on by the original human instructors, and 90.10% of these pieces of feedback obtained agreement from the human assessor."
Discussion¶
Related Claims¶
- GPT-4 demonstrated higher reliability of feedback than GPT-3.5, with higher F1 agreement with instructors on most assessment aspects — related
- GPT-4 outperformed GPT-3.5 and human instructors in producing feedback containing effective feedback components such as feeding-up and process-level information — related
- GPT models achieved high polarity agreement with instructors only for positive feedback on the Topic aspect, and could not reliably assess the other rubric aspects — related
- AI assessors show emerging reliability: acceptable AI-human rater agreement (Cohen's 0.78) in a 12-educator case study and human-comparable feedback helpfulness ratings with 45 learners — a broader claim this one bears on
Page metadata
| Field | Value |
|---|---|
type |
claim |
id |
human-assessor-agrees-with-gpt-feedback-on-uncommented-aspects |
title |
A human assessor agreed with the majority of GPT-generated feedback comments on assessment aspects that instructors had not commented on |
description |
A human assessor agreed with the majority of GPT-generated feedback comments on assessment aspects that instructors had not commented on |
status |
review |
evidence_strength |
moderate |
newest_evidence |
2024 |
recheck_by |
2027 |
generated |
by process:wiki-ingest on 2026-10-15 |
trust |
unverified — no one has confirmed this page's content |
sources |
1 — dai-2024 |
Frontmatter is the page's machine-readable record — see the Schema & Guide.