Due to technological advancements and changing requirements for higher education, there is a global trend toward returning to the assessment of genuine expression of students' thoughts. Educators are increasingly rejecting automated text detection software because true human expression is inherently imperfect and unique.
Dr. Philip Broster, a lecturer in the Professional Communication Unit at the School of Research Management at the University of Cape Town (UCT), studies how students cope with modern pressures related to technology and the professional environment. He notes that while it is sometimes obvious that students relied entirely on automated writing tools, if they use these tools as an aid, their usage becomes almost undetectable.
Brad Latilla Campbell, Country Manager at Crimson Education, shares this view, pointing out that international admissions committees are placing more trust in human judgment than in detection software. He emphasizes that strong academic performance only allows one to enter into a dialogue but does not guarantee an offer, as an ideal academic profile cannot compensate for an essay that seems constructed rather than sincere.
Abandoning Detection Software
The reality of these tools' limitations became evident in October 2025 when UCT discontinued the use of its primary detection tool, Turnitin, due to reliability issues. Sukaina Waldji, Director of the Centre for Innovation in Teaching and Learning at UCT, confirmed this decision, stating that the system had become unreliable as it could not keep up with the progress of generative AI. Turnitin began incorrectly flagging authentic student work while allowing automated text to pass.
Similarly, Stellenbosch University also ceased operating its automated text detection tool. Senior Advisor Dr. Haneli Adendorff believes that dependence on such systems is counterproductive to student-centered learning. Other institutions, including Nelson Mandela University, have adopted a similar stance, shifting the focus from algorithmic control to digital ethics education. Thus, Stellenbosch University is one of several South African institutions abandoning algorithmic text detection in favor of authentic student engagement and digital ethics.
Curriculum and AI
Dr. Philip Broster has integrated the ethical and practical aspects of AI use into his curriculum. He encourages personal narrative, an area where automated tools struggle to create original content but excel at polishing existing narratives. Broster explains that the assessment depends on how the student uses the tool, not whether they used it at all. He requires students to demonstrate a history of iterative software use while placing great weight on in-person exams and tests to ensure critical engagement.
By prioritizing personal storytelling, transparency of effort, and direct interaction, institutions like UCT and Stellenbosch affirm that learning cannot be measured solely by superficial perfection. Both local universities and global admissions committees agree that artificial intelligence can assist in research and structural refinement, but it cannot replicate genuine thought or true character.