Turnitin scores · updated 2026-11-09

What changed in Turnitin AI detection in 2026

The single blue highlight, Arabic support, Turnitin's own numbers on AI use, universities dropping detectors in Australia, the UK licence dispute, and what it all means for students.

2026 was the year AI detection became contested in public. Here is what changed, with dates, and what it means if you are submitting work this year.

August 2026: one blue highlight

Turnitin's AI report used to show two colours: blue for text likely written by AI, purple for text likely written by AI and then paraphrased by a tool. On 4 August 2026 Turnitin merged them. Everything the model considers likely AI is now one shade of blue. The model still looks for AI text that was run through paraphrasing and "humanizing" tools; it just stops showing that as a separate category. Reports generated before the change keep the old colours unless the file is resubmitted.

August 2026: Arabic

Turnitin added AI writing detection for Arabic submissions, joining English, Spanish, Japanese, German, Dutch and Turkish. Other languages still get a similarity report only.

Turnitin's own numbers

In February 2026 Turnitin said that since October 2025 about 15% of essay submissions had more than 80% AI-written text, up from 3% when the detector launched in 2023. In October 2026 it reported that students writing inside its own "Clarity" environment were about half as likely to outsource a paper almost entirely to AI. The company is investing in detection, not retreating from it.

Australia pulls back

In August 2026 news.com.au reported that major Australian universities had moved away from AI detection tools amid fears of false accusations. In September, the director of academic quality and integrity at Charles Sturt University wrote that the sector should stop using the tools altogether and redesign assessments instead. The University of Melbourne's student union documented students facing misconduct cases built around the score.

The UK licence dispute

In July 2026 Turnitin planned a new end-user licence covering how student submissions could be processed with AI tools. UK universities and sector bodies objected over data privacy, and Turnitin paused the change for twelve months. Universities including Birmingham and Queen Mary told students the service continues under the old terms for 2026/27. Turnitin then published a Trust Center stating it does not train a generative model on student essays, only a classifier.

Students and lawyers

The BBC reported in August 2026 on a St Andrews medical student accused over a dissertation he wrote himself. In October, the Financial Times reported that students facing AI accusations were increasingly hiring lawyers, one spending over £3,000 to clear their name with drafts and notes.

What it means for you

  • In the US and the UK, the score is still part of student life. Most schools use it, and the number from 20% up is visible to your professor.
  • The report changed shape, not meaning. A blue highlight now covers both AI text and AI-paraphrased text. Running an essay through a paraphrasing tool is not a safe fix.
  • Evidence beats argument. The cases that ended well in 2026 were won with drafts, version history and calm meetings. The evidence kit is built around that.
  • Check the file you will submit, when it is final. Model updates and file types change scores; a check from last term tells you little about this one.

Sources

Turnitin blog, 4 August 2026, "How Turnitin is simplifying AI detection"; Turnitin press releases, 24 February and 7 October 2026; Turnitin Trust Center, 1 October 2026; news.com.au, 13 August 2026; Charles Sturt University News, 24 September 2026; University of Birmingham student news, 5 October 2026; BBC News, 10 August 2026.

See your own score before you submit

Paste or upload your paper and get the Turnitin AI report back, the similarity report with it; the live turnaround is shown before you pay. If the score is high, you can fix it here.