Results from two decades of five-yearly plagiarism surveys: new insights into prevalence, understanding, attitudes, knowing vs naïve plagiarism, and generative artificial intelligence (genAI)

Historical trends in plagiarism are often estimated by comparing survey results from a diverse range of samples, institutions, and measures. However, in multi-institution, multi-method comparisons, changes over time are difficult to separate from differences in methods and context. We assessed self-reported prevalence, understanding, and perceived seriousness of seven forms of plagiarism in surveys of students at the same Australian university on five occasions over 20 years, each separated by 5 years (2004, 2009, 2014, 2019, and 2024). The new 2024 survey reported in this paper included: 310 students from the same institution as the previous surveys and 1867 from five other Australian universities, and an additional survey item assessing plagiarism from generative artificial intelligence (genAI). Results revealed a downward trend in plagiarism prevalence at the same university over time, accompanied by increased understanding and perceived seriousness of plagiarism. New insights include: the comparability of the original university’s results with the wider Australian sample; evidence that students plagiarise both naïvely and when they know the behaviour is plagiarism; plagiarism from generative AI may not have replaced traditional forms of plagiarism; and genAI detector programs have a negligible deterrent effect on plagiarising from genAI. These results suggest that although interventions may reduce plagiarism, it remains an ongoing educational and enforcement challenge. Not applicable.