News
Detecting AI-generated texts with Turnitin
WUR students are expected to avoid plagiarism by writing original texts for their assignments, reports and thesis. To detect plagiarism, WUR teachers use Turnitin. Turnitin has recently introduced a new AI detection module that detects if a text may have been written by AI writing tools, such as ChatGPT.
AI detection module in Turnitin
To assess student work, teachers need to know if the text is original and if AI writing tools may have been used. At WUR, Turnitin is available in Brightspace for teachers. Other staff can request a similarity check through the Copyright Information Point.
With the new AI detection module, the Turnitin software can detect whether a text was written by a person. Although the latest AI language models produce human-like texts, human-written texts are much less predictable and more inconsistent than AI texts. You can read more about Turnitin’s underlying technology and parameters in this FAQ on their website.
How to use this module?
If you‘re using Turnitin in Brightspace, the AI module is automatically activated. You’ll see the AI percentage next to the similarity percentage. The AI percentage indicates how much of the document may have been written by AI. It will also highlight any ‘suspect’ sentences that have been recognized as probably written by AI.
If you submit your work for a similarity check to the Copyright Information Point, the AI percentage will be available in the report you receive from the Point.
What to do with a high AI detection percentage in a student’s work?
The Turnitin analysis might give both false positives and false negatives. Be critical when reading the Turnitin report. We’d like to refer you to the Intranet page on artificial intelligence in education. There, you’ll find additional information as well as current and future guidelines on the use of AI.