Turnitin0 turnitin0

Turnitin0 - A trusted institution providing third-party Turnitin checking services to students worldwide.

Research report

Does Turnitin Flag Human-written ESL Essays as AI-generated?

Benchmark Experiment: Turnitin AI Detection Accuracy on Human-Written ESL Essays

James Peng

James Peng Turnitin0.com | Lead Editor

3 years of writing experiences on Turnitin AI detection and humanizer.

Overview

This benchmark evaluates Turnitin's AI detection accuracy on a corpus of 340 human-written ESL undergraduate essays from the CELL corpus. The goal is to determine whether Turnitin incorrectly flags authentic ESL writing as AI-generated. All essays are human-written, so any flagged text represents a false positive.

  • Overall word accuracy: 100.0% — Turnitin correctly classified all 263,329 words as human-written.
  • No false positives across any domain: Business (14,698 words), Education (40,890 words), Humanities (142,534 words), Psychology (43,155 words), and STEM (22,052 words) all achieved 100.0% word accuracy.
  • All 18 majors (e.g., Biology, Computer Science, English, Psychology) showed 100.0% word accuracy, with no misclassified words.
  • All word-length buckets (400, 800, 1200 words) also achieved 100.0% word accuracy, indicating consistent performance across essay lengths.
  • Dataset: 340 human-written ESL undergraduate essays from the CELL corpus, totaling 263,329 words (excluding references/bibliography).
  • Essays span five domains (Business, Education, Humanities, Psychology, STEM) and 18 majors, with word counts distributed across 400, 800, and 1200 word buckets.
  • Procedure: Each essay was submitted to Turnitin's AI detection system. The system's per-word predictions were compared against the ground truth (all human-written).
  • Metric: Word-level accuracy = (number of words correctly identified as human) / (total words) × 100.
  • 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.
  • Single corpus: Results are based on one ESL corpus (CELL) and may not generalize to other ESL populations or genres.
  • Limited demographic diversity: The CELL corpus primarily represents Chinese ESL learners; other L1 backgrounds may yield different results.

Linked datasets

Overall Results

  • Total articles: 340
  • Total words: 263,329
  • Correct words: 263,329
  • Word accuracy: 100.0%

Turnitin correctly identified all words as human-written, with zero false positives.

Results by Domain

  • Business: 27 articles, 14,698 words — 100.0% accuracy
  • Education: 47 articles, 40,890 words — 100.0% accuracy
  • Humanities: 186 articles, 142,534 words — 100.0% accuracy
  • Psychology: 42 articles, 43,155 words — 100.0% accuracy
  • STEM: 38 articles, 22,052 words — 100.0% accuracy

All domains achieved perfect word-level accuracy.

Results by Major

  • Biology: 1 article, 995 words — 100.0%
  • Business Administration: 2 articles, 1,584 words — 100.0%
  • Civil Engineering: 3 articles, 1,783 words — 100.0%
  • Communications: 84 articles, 56,961 words — 100.0%
  • Computer Science: 2 articles, 954 words — 100.0%
  • Criminal Justice: 7 articles, 3,772 words — 100.0%
  • Economics: 22 articles, 10,210 words — 100.0%
  • Education: 47 articles, 40,890 words — 100.0%
  • English: 79 articles, 66,503 words — 100.0%
  • Environmental Science: 24 articles, 14,365 words — 100.0%
  • Finance: 2 articles, 2,272 words — 100.0%
  • Health Sciences: 7 articles, 3,279 words — 100.0%
  • History: 1 article, 1,105 words — 100.0%
  • Information Technology: 1 article, 676 words — 100.0%
  • Marketing: 1 article, 632 words — 100.0%
  • Political Science: 1 article, 551 words — 100.0%
  • Psychology: 42 articles, 43,155 words — 100.0%
  • Sociology: 14 articles, 13,642 words — 100.0%

All 18 majors showed perfect word accuracy.

Results by Word Count Bucket

  • 400-word bucket: 125 articles, 57,028 words — 100.0% accuracy
  • 800-word bucket: 100 articles, 75,382 words — 100.0% accuracy
  • 1200-word bucket: 115 articles, 130,919 words — 100.0% accuracy

Accuracy remained perfect across all essay lengths.

Detection report screenshots