Universities stop using AI detection tools such as Turnitin
Vanderbilt University has decided to disable the AI detection tool provided by Turnitin due to concerns about its reliability and potential false positives. The tool was launched without providing detailed information on how it detects AI writing. Other companies offering AI detectors have either changed their business models or closed down. Instructors are advised to communicate with students about AI usage, establish guidelines, and consider revising assignments. They can look for inconsistencies in writing style and inaccuracies in sources to determine if a submitted text is AI-generated. AI tools are expected to become more prevalent, and resources are available to help instructors incorporate AI in teaching.
Vanderbilt University has decided to disable Turnitin’s AI detection tool after months of testing and consideration. Concerns over the tool’s reliability, potential false positives, and lack of transparency regarding its working mechanism prompted this decision, which aims to protect the best interests of students and faculty. When the AI detection tool was launched, many concerns arose due to the lack of notice, the absence of an option to disable the feature, and limited understanding of its workings.
Turnitin claimed a 1% false positive rate, but this could have led to around 750 student papers being wrongly identified as AI-generated out of the 75,000 submitted by Vanderbilt in 2022. Reports of false accusations of AI usage, including instances involving Turnitin, and the tendency of AI detectors to label non-native English speakers as AI writers have been widespread.
Furthermore, Turnitin has not provided detailed information about how it detects AI writing, leaving doubts about the accuracy of its claims. Other AI detection tool companies have shifted their business models or ceased operations altogether, raising concerns about the privacy and data usage policies of third-party detectors.
The difficulty of accurately identifying AI writing, especially as AI tools advance, has led Vanderbilt to conclude that AI detection software is not a practical solution. Moving forward, instructors are advised to communicate openly with students about AI usage, establish clear guidelines and expectations, encourage the use of disclosures and proper citations, and consider revising assignments to mitigate concerns. Instructors can compare submitted texts to students’ previous work and look for inconsistencies or inaccuracies as an indication of AI-generated content.
Resources are available to support instructors in adapting their courses to AI while maintaining academic integrity and leveraging the benefits of AI tools. As AI tools become more prevalent in everyday products and widely used software, educators must navigate the effective integration of AI into teaching practices.
Source: Vanderbilt University