Teaching, learning, and assessment with Generative AI
6. Detecting the use of Generative AI
Students believing the use of Generative AI can be detected encourages integrity. However, current detection approaches are only broadly indicative and may be evaded. Therefore, the challenges for markers of identifying generated work will increase (Sadasivan et al., 2023).
Using the Turnitin Generative AI indicator tool
At UQ the Turnitin Generative AI indicator tool is enabled. Other detection systems for AI writing are not recommended as they have not been reviewed to ensure they meet UQ's software policy (including privacy considerations). For staff who wish to use the Turnitin tool we recommend you use it cautiously as a ‘preview’ tool, and encourage you to take a supportive approach to help students use Generative AI appropriately.
A score on the Generative AI indicator alone is not evidence of academic misconduct, as the tool may create false positives and student use of Generative AI may still be appropriate in context. However, it should be a red flag and prompt you to investigate further for other indicators of misconduct in the assessment piece.