Short answer: Can Turnitin detect AI humanizers? Yes, far more often since 27 August 2025, when it added AI bypasser detection. Early testers saw some humanized texts jump from 0% to over 70% AI. It is not perfect, and it is trained for English only.
Can Turnitin detect AI humanizers since the 2025 update?
Yes, in many cases. On 27 August 2025, Turnitin added AI bypasser detection to its standard AI writing checks, aimed at text "modified by AI humanizer tools" (Turnitin). So how well can Turnitin detect AI humanizers now? It depends on the tool and the text.
Markers do not need to switch anything on. When AI detection is enabled, bypasser checks run automatically, and Turnitin's release notes say the "AI-generated only" category now includes AI text that may have been modified by a bypasser (iThenticate release notes).
What did early testers find?
Results were mixed across 6 tools in Tadhg Blommerde's YouTube test, reported by EdTech Innovation Hub. Some humanizers that once scored 0% were now flagged heavily, while others barely moved, so the question "can Turnitin detect AI humanizers?" has no single answer.
| Tool tested | Turnitin result after the update |
|---|---|
| StealthGPT | Rose from 0% to 72% likely AI |
| Groby | Rose from 0% to 67% likely AI |
| GPT Human | 31% likely AI |
| StealthWriter | Between 1% and 19%, shown as *% |
| Refrazy | Between 1% and 19%, shown as *% |
| Easy Essay | Stayed at 0% |
That is one small test, not a benchmark, and results shift with each model update.
Why can't Turnitin catch every humanizer?
Detection is a moving target, and 2 limits stop Turnitin catching everything. The bypasser feature is trained for English only. Each humanizer also changes its output over time, so can Turnitin detect AI humanizers next term? Nobody can say in advance. A Springer study found none of 14 detectors above 80% accuracy.
Turnitin also hides scores from 1% to 19% as *% because false positives are more likely in that range. So a humanized essay can land in the asterisk zone, and so can an essay a student wrote entirely alone. Our Turnitin false positive guide explains what the asterisk means.
Does Turnitin detect AI humanizers and paraphrasers the same way?
Not quite. Turnitin's AI report already had a label for AI text modified by a paraphrase tool before the 2025 change, and the bypasser update extends that to 1 more group: dedicated humanizers. So can Turnitin detect AI humanizers and paraphrasers alike? Broadly yes, and both rewrite styles now sit inside the AI writing percentage your marker sees.
Paraphrasers and humanizers work differently, which is why they score differently. Our guide to humanizers versus paraphrasers explains the gap, and does Turnitin detect QuillBot covers the best-known paraphraser on its own.
What should students do now?
In the test ETIH reported, Turnitin flagged 2 of 6 humanizers heavily (StealthGPT at 72%, Groby at 67%), so treat any tool as a draft helper and follow 2 rules. Check your course's AI policy first. Then keep your version history and drafts, and revise flagged sentences yourself.
Turnitin says its AI score should not be the sole basis for action against a student. If you are flagged on work you wrote, process evidence is your strongest defence. Our guide on how to prove you didn't use AI walks through version history, notes and what to say in a meeting.
Where does EvadeGPT fit in?
EvadeGPT stays honest about detection: it does not promise a pass, and its FAQ says "a human score or a bypass cannot be guaranteed". What it gives you is 3 humanizer models and a built-in AI detector on 1 screen, so you can see which sentences still read as machine-written.
The EvadeGPT humanizer keeps terms you add to its Keep words list, which helps when meaning matters more than a score. Use it to find weak sentences, then rewrite them yourself and keep your drafts.
FAQ
Can Turnitin detect AI humanizers?
Often, yes. Turnitin added AI bypasser detection on 27 August 2025, and early testers saw some humanized texts rise from 0% to over 70% AI.
Is Turnitin's humanizer detection perfect?
No. It is trained for English only, results vary by tool and text, and scores from 1% to 19% show only as *%.
Do markers have to turn on bypasser detection?
No. When AI writing detection is enabled, Turnitin checks for bypasser tools automatically.
If Turnitin can detect AI humanizers, does EvadeGPT still help?
It helps you find weak sentences, but its own FAQ says a human score or a bypass cannot be guaranteed, so check your draft and keep your version history.
Can Turnitin detect AI humanizers in other languages?
Not yet. Turnitin says the bypasser detection is trained and tested on English only.
