South African private tertiary institution Stadio has denied claims that it has begun using artificial intelligence to mark student papers, even as it has formalized a structured framework for student AI use that mandates the technology in some assignments.
A MyBroadband reader who had previously received payment for assisting with the marking of Stadio assignments alleged that the institution had begun reducing human marking workloads with AI tools. A lecturer at the institution who had outsourced some marking to the reader said students were now expected to upload more of their assignments for online marking, with some assignments requiring only that lecturers upload memorandums while Stadio’s systems handled the rest. The reader said the number of assignments they received from the lecturer to mark had dropped from approximately four per year to one or two.
Esther Venter, dean for teaching, learning and student success at Stadio Higher Education, rejected the allegation. “Marking is conducted by appropriately qualified academic staff, in line with our institutional policies and quality assurance processes,” she said. Venter added that any potential future use of AI tools in marking would be carefully evaluated to ensure alignment with academic standards, fairness and regulatory requirements. “At present, all marking decisions remain the responsibility of human academics, and established processes are in place to ensure quality and consistency.”
The proliferation of generative AI tools has been a hotly debated subject in higher education since ChatGPT’s launch in November 2022. Some institutions initially attempted to restrict the technology, particularly student use in assignments, often relying on AI detection tools that have been widely criticized as unreliable.
A 2025 presentation by Stadio’s LLB Department head, Jacques Nieuwoudt, argued that AI detection tools as a stand-alone approach were a dead end. Nieuwoudt also said AI detection widens the digital divide, as students from more privileged backgrounds may find it easier to manipulate outputs to avoid detection.
Stadio has since adopted a structured framework for AI use, recently announcing a “traffic light” model that not only permits student AI use in certain assignments but explicitly requires it in others. The model classifies assignments as follows: red, where AI use is not permitted, for assessments designed to build independent skills such as critical thinking, communication and problem-solving; amber, where AI use is optional, for tasks requiring brainstorming and editing; and green, where AI use is required, for repetitive tasks or idea generation aimed at boosting productivity.
Jolanda Morkel, Stadio’s head of instructional design and a senior research academic, said the institution recognized that AI was not going away and that its use in education should not be feared. AI skills, she argued, were becoming as important as reading, writing and numeracy. “In the workplace, there will be tasks where AI is not appropriate, tasks where it may be helpful, and tasks where it is indispensable,” she said. “AI is one of the defining technologies of our time. By embracing it as a partner in education, we are ensuring that our graduates are prepared not just to adapt, but to thrive.”
