Today, Turnitin, a leading provider of technology solutions in academic integrity, announced its AI paraphrasing detection feature. Integrated within Turnitin’s AI writing capabilities, the feature helps educators and publishers identify when AI paraphrasing tools may have been used to modify AI-generated text to avoid detection.
With the proliferation of generative AI tools, ensuring original critical thinking in education and research writing has become increasingly challenging. Writers, including students and researchers, may use AI paraphrasing to modify AI-generated text created by large language models, leading educators and publishers to believe that the work is original. A recent study conducted by Tyton Partners* found that students continue to lead educators and administrators in their use of generative AI tools. Fifty-nine percent of students are regular users, monthly or more, compared to approximately 40 percent of educators and administrators.
Turnitin’s AI paraphrasing detection feature helps address this issue by identifying instances where AI paraphrasing may have been used to modify AI-generated text. The AI paraphrasing detection feature enables educators and researchers to initiate constructive conversations with writers about the importance of ethical writing practices and proper paraphrasing techniques.
“AI has incredible potential to benefit education but writers need to be transparent in how they use Generative AI,” said Annie Chechitelli, chief product officer, Turnitin. “Applying AI paraphrasing to AI-generated content is likely a sign that the writer is trying to hide their use of AI, which is the opposite of transparency. Turnitin provides the tools and insights to preserve the value of original work.”
In early 2023, Turnitin launched its AI writing detector and report, leveraging over 25 years of experience in understanding and safeguarding academic writing. Of the over 200 million papers reviewed* since the launch, Turnitin data shows:
-
Over 22 million papers (approximately 11 percent of over 200 million) have at least 20 percent AI writing present**
-
Over six million papers (approximately 3 percent of over 200 million) have at least 80 percent AI writing present**
