For educators wringing their hands over the threat of students using ChatGPT to cheat on their assignments, there’s good news. Turnitin has developed technology that can detect AI writing and can identify submissions generated by ChatGPT with high accuracy. Their originality check, Turnitin Originality, can evaluate the source of text in a submission and highlight any AI-generated content. They have also added an AI writing indicator to their Similarity Report, allowing professors to identify any sentences that may be AI-generated.

This new feature is how to get around turn it in into the existing similarity report workflow. The AI writing indicator displays the percentage of prose sentences within a long-form writing format (such as an essay) that Turnitin suspects may be AI-generated. However, it is important to note that this model is not designed to distinguish between speech-to-text and other forms of AI-generated text, such as poetry, scripts, code, ordered/unordered lists, or tables.

When I put this into Originality it shows as blatant ChatGPT:

While this new capability is a step in the right direction, it is also vital for educators to recognize that these tools are not foolproof and may falsely flag student work. This could erode student trust in educators and lead to distrust of the integrity of assessment processes, as well as strained relationships between faculty and students.

Educators should encourage their students to continue building their own writing skills and use these tools as a supplement, rather than a substitute for their written assignments. They should also instruct their students to properly cite any ideas or information they’ve gathered from outside sources and provide proper attribution when using these tools in their assignments.