As generative AI writing tools like ChatGPT can turnitin detect ai writing to rise in popularity, educators are concerned students could use the technology to cheat on assignments. Fortunately, Turnitin’s plagiarism detection software is evolving to help prevent this issue by identifying AI-generated text in student papers. Educators at 10,700 institutions and more than 2.1 million educators will be able to access the new AI detection capabilities today, according to Turnitin.
The company’s new tool detects AI-generated writing, including that created with the exploding generative AI tool, with 98% accuracy. The company began developing the software nearly two years ago, and its technology is continually improving. It has also worked to ensure a low false positive rate so that teachers don’t unfairly accuse innocent students of plagiarism.
Evaluating the Authenticity: A Guide to Checking AI-Generated Content for Accuracy
Turnitin’s AI detection feature is activated when an educator submits a submission that contains AI-generated writing. The submission is then compared against its database of existing content to identify potential plagiarism. If the paper matches, a flag is placed in the report to indicate it may contain AI-generated text. Educators can then decide how to respond to the AI-generated text in their submission.
Turnitin’s AI writing detection tool is currently based on OpenAI’s GPT family of large language models. It will soon be updated to include GPT-4, and the company plans to future-proof its software for advanced AI writing tools that may emerge in the coming years. The new feature can be accessed in the AI Indicator section of a submission report. A blue percentage will display the amount of qualifying text in a submission that Turnitin’s model determines was generated by AI.