Higher education is in a state of panic over Artificial Intelligence writing essays. But according to Professor Fulufhelo Nemavhola (Deputy Vice-Chancellor at Durban University of Technology), academia is worrying about the wrong problem.
The crisis isn’t academic cheating; it’s graduate capability. AI didn’t break university assessments—it merely pulled back the curtain on decades of “assessment theatre.”
For generations, universities have suffered from the fluency fallacy: the dangerous assumption that polished language, flawless referencing, and scholarly structure equate to genuine human understanding. Today, Generative AI can mimic that academic confidence at scale, satisfying grading rubrics perfectly while remaining completely devoid of actual thought.
The Stakes for South Africa: We Can’t Afford Empty Fluency
In a nation wrestling with severe youth unemployment, infrastructure decay, water insecurity, and energy instability, this flaw in the higher education system has devastating, real-world consequences. South Africa desperately needs problem-solving universities that test actual competency, not just elite writing skills.
Banning AI or hiding behind detection software is a losing, defensive strategy. AI is already deeply embedded in modern classrooms, research, and corporate boardrooms. The goal should not be to make assessments “AI-proof,” but to make genuine learning visible.
The Solution: Moving to Judgement-Rich Assessment Ecologies
To survive and thrive in the machine age, universities must pivot from submission-based grading to judgement-rich evaluation. Writing is still vital, but a written report can no longer stand alone as unquestioned proof of intelligence.
Instead, higher education must implement an assessment ecology that demands live human accountability:
- Engineering: Don’t just grade the blueprint. Make the student orally defend their load assumptions, material constraints, and failure scenarios.
- Coding: Don’t just grade the code. Require the student to live-debug errors and explain their architectural choices.
- Public Policy: Don’t just grade the paper. Force the student to verbally debate the ethical trade-offs, risks, and economic consequences of their choices.
Embracing Accountable AI Usage
The future graduate will not be someone who has never used AI—that is an unrealistic expectation in the 2026 workforce. The future graduate must be someone who can use AI critically, ethically, and intelligently.
Students should be required to declare their AI workflows openly: what outputs they accepted, what they rejected, how they verified the facts, and exactly where human judgement took over. True education means moving away from a culture of concealment and toward a culture of transparent, accountable capability.
