THE WIDESPREAD AVAILABILITY and use of generative artificial intelligence (GenAI) tools, such as DeepSeek and ChatGPT, present significant challenges to gauging the authenticity and originality of higher education assessments. Despite the emergent challenges, the rise of GenAI does not signal the end of assessments altogether. Rather, it necessitates their evolution to determine the degree to which authentic learning takes place. What are the implications for the institutional integrity of higher education, and what types of assessment can ensure that students are still learning, even as the usage of these tools becomes inevitable?
HIGHER EDUCATION IN THE TIME OF AI
The prevalence of GenAI tools can potentially undermine the integrity of higher education itself. Whether gauged through written assignments, essays or other assessments, it is increasingly difficult to ascertain if students are actually doing the hard work required of them. Traditional assignments are text-based: they require students to read texts and produce lengthy reflective essays, or conduct and report on specialised research. In short, they demand time, critical ability and sustained attention. But they can now be completed with relative ease using GenAI tools, which essentially offer an alternative to original written engagement. As a result, such assessments can be easily gamed, and can no longer claim to reliably measure student learning or capability.
Indeed, the extent of GenAI usage is nearly impossible to measure accurately. AI-detection functions do exist on learn-ing management platforms, but these can only provide estimates on the likelihood of AI use. Such indeterminacy hardly constitutes an objective platform for properly determining whether or not students are applying themselves. When the validity of assessment systems is compromised, the ability of educators to determine quality and appropriately reward genuine effort, critical thinking and academic excellence is diminished.
