Do AI Detectors Like Turnitin Actually Work? What the Research Says
Short answer: AI detectors are far less reliable than most people think. They can flag human writing as AI, miss AI writing that has been lightly edited, and even the companies behind them say a score should never be used as proof on its own. Here is what the research and real cases show.
What AI detectors do
Tools like Turnitin's AI writing indicator, GPTZero and others estimate how likely it is that text was written by AI. They look at patterns such as how predictable the word choices are. The result is a probability, not a fact.
What the research shows
- No detector is close to perfect. A 2023 study by Weber-Wulff and colleagues tested 14 detectors, including Turnitin, and none reached 80% overall accuracy.
- Human writing gets flagged. A 2023 comparison of 16 detectors found false-positive rates on human essays ranging from 0% to 31%, with Turnitin doing better than most.
- Non-native English writers are flagged far more. A Stanford-led study found seven popular detectors flagged an average of 61.3% of essays by non-native English writers as AI-generated.
- Light editing fools them. A 2024 study found that small changes to AI text cut detection accuracy sharply across the tools tested.
- Even OpenAI gave up on its own detector. It withdrew its AI text classifier in July 2023 because of low accuracy.
Source: research summary in Mount Royal University's AI detection advisory.
What Turnitin itself says
Turnitin's own guidance says its AI score should not be the only basis for action against a student. It also hides exact scores in the low range because false positives are more common there.
Universities are turning it off
Several universities have stopped using or declined Turnitin's AI detection feature, including Vanderbilt, Waterloo, UBC and the Australian National University. Washington State University reported that a third of its AI misconduct cases from 2023 to 2025 were dismissed when detection was the only evidence.
If your own work gets flagged
- Stay calm. A flag is not proof.
- Gather your evidence: outlines, notes, earlier drafts, and Google Docs or Word version history.
- Ask what the evidence is. Politely ask your teacher whether the score is the only reason.
- Offer to explain your work. Talking through your ideas and sources is strong proof you wrote it.
- Know your school's process. Most have an appeal or review procedure.
How to protect yourself from day one
- Write in Google Docs or Word with version history on.
- Keep your notes and outlines.
- Follow your course's AI rules and say openly if you used AI for brainstorming or feedback.
- Use AI to learn and get feedback, not to write work you submit as your own.
Looking for AI tools that help you study the right way? See our guides for students.