Turnitin uses its own proprietary AI writing detection technology designed specifically to identify text generated by AI tools such as ChatGPT. This AI detector is integrated into Turnitin's Similarity Report and is specialized for student writing, leveraging Turnitin's 25 years of experience in academic integrity
. The detection method analyzes patterns such as the probability of word sequences and compares them to typical AI-generated text, distinguishing it from human writing by identifying less idiosyncratic and more statistically probable word choices typical of AI
. Turnitin claims its AI detection tool is about 98% accurate and aims to balance detection with minimizing false positives, accepting that it may miss about 15% of AI-generated content to keep false positives under 1%
. The AI writing detection results appear as a percentage indicating the amount of AI-generated text detected, with scores below 20% flagged with an asterisk to caution about reliability due to higher false positive rates in that range
. This feature is only visible to instructors and not to students
. In summary, Turnitin uses a proprietary AI detection model integrated into its plagiarism and originality checking system, focused on student writing, with high accuracy and built-in safeguards to reduce false positives
