Research report
Does Turnitin Flag Human-Written Research Papers from the PLOS Corpus as AI-generated?
Benchmark Experiment: Turnitin's AI Detection Accuracy on 504 Human-Written Graduate-Level PLOS Essays
Cite this research
In a benchmark of 504 human-written graduate-level essays from the PLOS corpus (135,712 words), Turnitin achieved 100.0% word-level accuracy, correctly classifying every word as human-written across 18 majors and four domains.
According to Turnitin0 Research, the AI detection system correctly identified all 135,712 words from 504 human-written PLOS essays as human-authored, achieving perfect word-level accuracy across all 18 majors and four domains.
- 504 articles tested
- Checker: Turnitin
- Corpus: PLOS
- 1 linked dataset
- Detection screenshots available
- Published July 28, 2026
Peng, J. (2026). Does Turnitin Flag Human-Written Research Papers from the PLOS Corpus as AI-generated?. Turnitin0 Research. https://www.turnitin0.com/research/reports/does-turnitin-flag-human-written-research-papers-from-the-plos-corpus-as-ai-generated
Licensed under CC0 1.0 Universal. You may copy, modify, distribute, and cite this research—including text, figures, tables, and datasets—without asking permission.
Overview
This benchmark evaluates whether Turnitin's AI detection system incorrectly flags human-written graduate-level research papers as AI-generated. The dataset comprises 504 essays from the PLOS corpus, totaling 135,712 words, spanning 18 majors across Business, Humanities, Psychology, and STEM domains. All essays are non-ESL and fall within the 400–800 word range. The key metric is word-level accuracy: the proportion of words correctly classified as human-written.
- Overall word accuracy: 100.0% — Turnitin correctly classified all 135,712 words as human-written across 504 articles.
- No false positives: The system did not flag any human-written text as AI-generated at the word level.
- Consistent performance across domains:
- Business: 1,291 words correct (100.0%)
- Humanities: 2,942 words correct (100.0%)
- Psychology: 10,870 words correct (100.0%)
- STEM: 120,609 words correct (100.0%)
- Uniform accuracy across majors: All 18 majors, from Biology (72,983 words) to Sociology (830 words), achieved 100.0% word accuracy.
- Word count buckets: Both the 400-word bucket (134,954 words) and the 800-word bucket (758 words) showed perfect accuracy.
- Dataset: 504 human-written graduate-level essays from the PLOS corpus, totaling 135,712 words (excluding references/bibliographies).
- Essay characteristics:
- Length: 400–800 words per essay.
- Non-ESL (English as a second language) writers.
- 18 majors including Biology, Computer Science, Psychology, Economics, and more.
- Detection tool: Turnitin's AI detection system.
- Evaluation metric: Word-level accuracy — the percentage of words correctly identified as human-written (not AI-generated).
- Analysis dimensions:
- Overall (all articles)
- By domain (Business, Humanities, Psychology, STEM)
- By major (18 categories)
- By word count bucket (400 and 800 words)
- Scoring unit: accuracy is measured in words. First identify correct paragraphs under the expectation below, then sum the words in those paragraphs for the correct-word count; total words are summed over all counted paragraphs.
- Correct paragraph: a counted paragraph that was not flagged as AI. Correct words are the words in those paragraphs.
- References excluded: from a References/Bibliography-style heading onward, those paragraphs are left out of word counts.
- Limited word count range: All essays are 400–800 words; performance on longer or shorter texts is not assessed.
- Total articles: 504
- Total words: 135,712
- Correct words: 135,712
- Word accuracy: 100.0%
- Turnitin correctly classified every word as human-written, with no false positives.
- Business: 6 articles, 1,291 words — 100.0% accuracy
- Humanities: 11 articles, 2,942 words — 100.0% accuracy
- Psychology: 40 articles, 10,870 words — 100.0% accuracy
- STEM: 447 articles, 120,609 words — 100.0% accuracy
- All domains achieved perfect word-level accuracy.
- Biology: 280 articles, 72,983 words — 100.0%
- Biomedical Engineering: 36 articles, 9,784 words — 100.0%
- Chemistry: 1 article, 309 words — 100.0%
- Civil Engineering: 1 article, 271 words — 100.0%
- Communications: 4 articles, 1,077 words — 100.0%
- Computer Science: 12 articles, 2,914 words — 100.0%
- Criminal Justice: 1 article, 162 words — 100.0%
- Data Science: 17 articles, 4,026 words — 100.0%
- Economics: 6 articles, 1,291 words — 100.0%
- Environmental Science: 15 articles, 3,915 words — 100.0%
- Health Sciences: 80 articles, 25,137 words — 100.0%
- History: 1 article, 285 words — 100.0%
- Mechanical Engineering: 1 article, 223 words — 100.0%
- Nursing: 2 articles, 543 words — 100.0%
- Physics: 2 articles, 504 words — 100.0%
- Political Science: 2 articles, 588 words — 100.0%
- Psychology: 40 articles, 10,870 words — 100.0%
- Sociology: 3 articles, 830 words — 100.0%
- All 18 majors showed perfect word accuracy.
- 400-word bucket: 503 articles, 134,954 words — 100.0% accuracy
- 800-word bucket: 1 article, 758 words — 100.0% accuracy
- Both word count ranges achieved perfect word-level accuracy.
Detection report screenshots
Turnitin marks suspected AIGC text in cyan.