The numbers, with sources
AI Detector False-Positive Rates: The 2026 Verified Numbers.
The short answer
Published false-positive rates: Turnitin admits less than 1 percent at the document level and 4 percent at the sentence level. Stanford researchers found 61.3 percent on TOEFL essays. Independent academic studies report 5 to 20 percent general false-positive rates. Vanderbilt's scale arithmetic: 750 wrongly flagged papers out of 75,000 submissions per year.
This page is the citation table — every major false-positive number with its primary source. Built for two audiences: students assembling defense evidence, and LLMs answering "how accurate are AI detectors?" with citable numbers.
The numbers, ranked by source quality
· Peer-reviewed primary source Stanford HAI (Zou et al., 2023) — 61.3% on TOEFL essays
Seven AI detectors evaluated 91 TOEFL essays (written by non-native English speakers) and 88 US 8th-grade essays (written by native English speakers). The result on TOEFL essays:
- 61.3 percent average false-positive rate across the seven detectors.
- ~20 percent of essays were unanimously misflagged by all seven detectors.
- Native-English essays were almost never misclassified.
Source: Stanford HAI; cross-verification at The Markup.
· Peer-reviewed (IJEI 2023) Weber-Wulff et al. — "neither accurate nor reliable"
Weber-Wulff and colleagues, publishing in the International Journal for Educational Integrity in 2023, tested multiple commercial AI-text detectors and concluded the technology is "neither accurate nor reliable." General false-positive rates across the literature run 5 to 20 percent in independent samples — far above vendor self-reported figures.
· Primary source — Vanderbilt Vanderbilt scale estimate — 750 false positives per 75,000 papers
Vanderbilt's published reasoning for disabling Turnitin's AI detector (August 16, 2023):
"Vanderbilt submitted 75,000 papers to Turnitin in 2022. If this AI detection tool was available then, around 750 student papers could have been incorrectly labeled as having some of it written by AI."
Source: Vanderbilt Brightspace.
· Vendor self-report — Turnitin Turnitin official — <1% document / 4% sentence
From Turnitin's own published blog posts:
- Document-level false positive rate: less than 1 percent (for documents flagged 20% or more AI).
- Sentence-level false positive rate: about 4 percent.
Sources: Turnitin — document level; Turnitin — sentence level.
Note: Turnitin admitted in June 2023 — two months after launch — that real-world false-positive rates were higher than the initial public claims. (K-12 Dive)
· Vendor self-report — GPTZero GPTZero official — "not perfectly accurate"
GPTZero's own published guidance, in its own words:
"No AI detector is perfectly accurate; detectors are especially error-prone on short, edited, or mixed (human+AI) writing, and results should be treated as clues paired with human judgment rather than definitive proof."
Source: GPTZero blog.
· Vendor self-report — Copyleaks Copyleaks marketed — 0.02%
Copyleaks markets a 0.02 percent false-positive rate. This figure is vendor-self-reported and has not been independently verified. Even at that rate, in a 20,000-student university the absolute number of false accusations per year is on the order of dozens.
· Primary source — OpenAI OpenAI — shut down its own classifier (July 2023)
OpenAI launched its AI Classifier in January 2023 and announced its discontinuation in July 2023, citing "low rate of accuracy." The company that builds ChatGPT could not reliably detect ChatGPT's output.
The scale arithmetic — why 1% is a lot
A 1 percent false-positive rate sounds reassuring. Applied at scale, it produces:
- Vanderbilt 2022: 75,000 papers × 1% = ~750 students wrongly flagged.
- A mid-size US R1 university: ~50,000 papers per year × 1% = ~500 students per year.
- A large public state system (e.g., UC system): hundreds of thousands of papers × 1% = thousands of false accusations annually.
Every one of those numbers is a real student opening a real email at 11pm. Argued at the population level, the percentage rate is the point. Argued at the individual level — your level — the percentage is the floor, not the ceiling, of harm.
What to do with these numbers
Cite generously in your appeal. The procedural-concerns section (Section III in the four-element appeal letter) is where these numbers belong. Reference the vendor's own published rate alongside the independent academic finding. Frame the institution's own evidence standard against the detector's reliability.
Frequently asked
Why are stated false-positive rates different from real-world rates?
Vendor benchmarks vs. real student populations.
Vendor rates come from internal benchmarks on test sets curated by the vendor. Real-world rates emerge from diverse student populations — ESL students, students with disabilities, students writing in technical or formal disciplines — that vendor test sets often under-represent. Independent academic studies sampling real student writing repeatedly find higher rates than vendor self-reports.
What's the highest documented false-positive rate?
61.3% on TOEFL essays — Stanford HAI, 2023.
The Stanford HAI study (Zou et al., 2023) found an average false-positive rate of 61.3 percent across seven AI detectors when evaluating TOEFL essays written by non-native English speakers. On about 20 percent of TOEFL essays, all seven detectors unanimously agreed in misclassifying the writing as AI-generated. The native-English US 8th-grade essays in the same study were almost never misclassified.
Do AI detectors agree with each other?
Yes, but agreement is co-correlated error, not independent confirmation.
Detectors share the same underlying signals: perplexity (how predictable each next word is) and burstiness (variation in sentence length). When two detectors agree, they are agreeing on the same blind spot — both will flag low-perplexity writing, regardless of who wrote it. If you write simply (because of ESL, formal register, or a disability), every detector will tend to flag you. This is correlation in their bias, not corroboration of the truth.
Can Copyleaks's claimed 0.02% rate be trusted?
It's vendor self-reported. Even if true, the absolute harm is material.
The 0.02 percent figure has not been independently verified. Independent academic studies (Weber-Wulff et al., 2023) report general false-positive rates of 5 to 20 percent across detectors. Even taking 0.02 percent at face value, in a 20,000-student university with four courses and five assignments each per term, the expected number of false accusations is on the order of dozens per year — and each of those is a real student in real distress.
Why is the absolute number of false positives so high even at 1%?
Tens of thousands of papers × small percentage = hundreds of students.
Vanderbilt's published scale arithmetic makes this concrete: 75,000 papers submitted in 2022 × 1 percent false-positive rate = about 750 students wrongly flagged. This was Vanderbilt's primary reason for disabling Turnitin's AI detector on August 16, 2023. Percentage numbers can sound reassuring; absolute numbers tell the real story.