![Out of 65 million papers, more than 2.1 million were detected as having at least 80% AI writing [File]](https://th-i.thgim.com/public/sci-tech/technology/4g50hr/article65085043.ece/alternates/LANDSCAPE_1200/polandcyber.jpg "Out of 65 million papers, more than 2.1 million were detected as having at least 80% AI writing [File]")
Plagiarism and similarity checker Turnitin said on Tuesday it has reviewed more than 65 million documents since the April launch of its feature, which detects similarities to AI-generated text.
Out of over 65 million papers, more than 2.1 million were detected as having at least 80% AI writing, while around 6.7 million had more than 20% AI writing. This comprises around 3.3% and 10.3% of the reviewed papers, respectively.
Turnitin stated these statistics were a sign that generative AI had become a part of classrooms, regardless of educator policies surrounding the new technology.
“In just over three months, Turnitin’s new AI detection feature is being used widely and it is giving educators data they have asked for since ChatGPT first became free to the public in November 2022,” said chief product officer at Turnitin, Annie Chechitelli.
(For top technology news of the day, subscribe to our tech newsletter Today’s Cache)
“Since the release, nearly 98 percent of Turnitin institutions have AI writing detection enabled within their Turnitin workflow in at least one of their active accounts,” she added.
However, Turnitin has also faced claims that it mistakenly classified original human writing as being largely AI-generated.
