If you have looked at a Turnitin AI writing report recently, you may have noticed something different: the purple highlights are gone, and everything flagged as likely AI now shows in a single shade of blue.
That is not a glitch. Turnitin announced the change on 4 August 2026 in a post by its Chief Product Officer, Annie Chechitelli. Here is what students need to know.
What changed
The old report sorted likely AI content into two categories with two colours. The new one merges them:
| Type of content | Before August 2026 | Now |
|---|---|---|
| Likely AI-generated text | Blue | Blue |
| Likely AI-paraphrased or bypasser-modified text | Purple | Blue |
The report no longer distinguishes between text copied straight from a large language model (LLM) and text that was generated and then run through an AI paraphraser or a so-called "bypasser" or "humanizer" tool. If the system judges it likely AI-produced, it gets the same blue highlight.
Why Turnitin changed it
Turnitin's explanation is that AI rewriting is now built into many everyday writing platforms, so the line between "AI-generated" and "AI-modified" has blurred. A student can have one AI draft an assignment, then refine it with another tool until the original signals fade. In that world, showing the two categories separately invites overinterpretation, so the report was simplified into one view.
Does this mean detection got weaker?
No. Turnitin states that its detection has not been reduced and that it still detects AI text that has been paraphrased or run through bypassing tools. The change is in how results are displayed, not in what is detected.
The practical takeaway for students: running AI text through a "humanizer" does not make it safe. That text used to appear in purple; now it appears in the same blue as directly generated text.
One caveat from Turnitin's footnote: paraphraser and bypasser detection currently applies to English submissions only.
Do old reports update automatically?
No. AI writing reports generated before the change keep the two-colour view. A file has to be resubmitted to get a report in the new format — which may explain why an older report you have looks different from a newer one your instructor sees.
What has not changed
- The AI score is not proof of misconduct — Turnitin's consistent position is that AI writing scores should start a conversation between instructor and student, and should never be the sole basis for an integrity decision.
- Low results still show an asterisk — when detection falls in the 1–19% range the report shows
*%and highlights nothing. See What does *% mean in Turnitin AI detection? - Students usually cannot see the AI score — according to Turnitin's August 2026 guidance, the AI writing score is currently visible only to instructors and administrators.
What students should do
- Read your course's AI policy first — some courses allow AI for brainstorming or proofreading, others ban it entirely. It is usually in the course outline or assignment brief.
- Watch out for built-in AI rewriting — "rewrite this sentence" or "make it more formal" features in word processors and grammar tools are exactly the AI modification Turnitin describes. Even if the argument is entirely yours, AI-rewritten sentences can be highlighted.
- Keep a record of your writing — notes, outlines, drafts and document version history are your strongest evidence if your work is ever questioned.
- Declare honestly — if your course allows AI, state which tools you used and for which parts, as required.
Check before you submit
Because students generally cannot see their own AI score, it helps to know what a report will show before you submit for real. TurnCheckHK uses an international Turnitin instructor account set to No Repository, and reports include both similarity and AI writing detection results. See pricing or create a free account.

