Since the launch of ChatGPT in late 2022, concerns regarding university assessments have surged. The critical question emerges: Are we truly testing students’ learning, or merely their proficiency in using generative AI?
In response to these uncertainties, universities and educators have been proactively reforming assessments to preserve the integrity of degrees as valid indicators of students’ knowledge and skills. However, navigating this landscape poses an ongoing challenge for institutions worldwide.
AI Doesn’t Always Equal Cheating
In July, Turnitin, a prominent US software company, reported a striking statistic: nearly 54% of university submissions from Australian students, analyzed between October 2025 and April 2026, displayed some form of AI utilization. Alarmingly, about 10% of these submissions incorporated over 80% AI-generated content. However, the mere presence of AI in academic submissions does not automatically imply that students are engaging in academic dishonesty.
Many students are leveraging AI as a tool to enhance their work rather than replace their critical thinking. For instance, they might employ AI to brainstorm ideas, refine sentences, edit drafts, or solicit feedback on arguments. This nuanced application of technology emphasizes the need for educators to differentiate between constructive use of AI and outright cheating. A recent study conducted with 24 students in 2024 revealed that over half utilized AI to assist with reading academic texts, employing it to summarize complex materials and clarify their interpretations.
This article is part of a series on how artificial intelligence (AI) is reshaping higher education.
We take a deeper look into how AI has disrupted degrees, how universities are responding to these new challenges, and what the technology means for the sector’s future. You can find more stories here.
Rules Aren’t Always Clear
Despite the growing awareness of AI’s presence in academic settings, research indicates that the distinction between acceptable and unacceptable use of AI is often blurry for both students and instructors. For instance, can improving a draft with AI be limited to correcting grammatical errors, or does it extend to reworking entire arguments? This ambiguity complicates the establishment of clear guidelines for both students and educators.
Another challenge arises from the fact that the parameters for appropriate AI use can vary significantly between assessments. Even within the same assessment, individual students might navigate these guidelines differently based on their comprehension and skills.
How Are Universities Responding?
Australian universities have a regulatory imperative to ensure their assessments accurately reflect that students have grasped the requisite knowledge and skills of their courses. However, their responses and strategies have varied widely. One notable approach gaining traction is the University of Sydney’s “two-lane approach,” which has been adopted by multiple institutions across Australia.
This two-lane strategy categorizes assessments into two distinct types. The first lane, labeled “secure,” mandates that students showcase their capabilities and knowledge without the assistance of AI. Typically, these assessments occur in person, such as traditional exams or interactive oral assessments, allowing educators to monitor student performance directly.
The second lane, referred to as “open,” permits students to utilize AI freely, commonly seen in online quizzes and essay assignments. Since assessing AI’s inappropriate application in these contexts proves challenging, “open” assessments usually aim for feedback and learning rather than as definitive indicators of student proficiency.
Universities Are Still Struggling to Keep Up
While increasing reliance on in-person assessments may seem like a feasible solution, it is not without its limitations. In-person evaluations still do not fully mitigate the risk of AI influence. Innovative wearable technologies, such as AI-enabled smart glasses, pose a significant challenge, as they can provide covert exam assistance without easy detection by invigilators. Reports have emerged of students using such devices during exams in countries like New Zealand, Taiwan, and South Korea, creating additional hurdles for institutions trying to maintain academic integrity.
Are Universities on the Right Track?
Our research illustrates AI and assessment as a “wicked problem”—a complex issue that lacks a straightforward solution. The crux of the struggle lies in the uncertainty surrounding whether students are using AI as a support for their learning or to undermine faculty assessments. It is crucial for universities to gather credible evidence regarding what graduates are genuinely capable of, yet approaching assessment merely as a means of controlling AI use may unintentionally shift the focus from effective learning to the efficacy of surveillance measures.
Instead of solely evaluating whether a new assessment method functions in an AI-centric world, it may be beneficial to consider which problems these methods resolve and which new challenges they generate. A focus on preventing AI interference may yield immediate solutions, such as increased supervision and stricter controls. However, if the objective is to equip students for effective AI engagement in their future careers, such restrictions could hinder the very learning outcomes universities aim to promote. This suggests that what constitutes a “solution” in one context might yield different implications in another.
For universities to advance in this arena, they must delve deep into identifying which assessment reform strategies are effective—and which are not—across various disciplines and contexts, rather than relying on a singular “perfect” solution universally applicable throughout an institution.
Inspired by: Source
