Why UCT Pulled the Plug on AI Detection

In a landmark decision, the University of Cape Town (UCT) has discontinued AI detection tools like Turnitin’s AI Score, calling them "fundamentally flawed" for academic integrity. Sukaina Walji, Director of UCT’s Centre for Innovation in Learning and Teaching, revealed the move stems from the tools’ unreliability—flagging false positives (human work mislabeled as AI) and missing AI-generated content altogether.

"These tools create mistrust between staff and students," Walji emphasized, noting their inability to keep pace with increasingly sophisticated AI writing. The decision aligns with global skepticism; institutions like Vanderbilt and MIT have similarly questioned AI detectors’ efficacy.

The New Blueprint: Observable Assessments & AI-Enhanced Learning

Rather than policing AI use, UCT is overhauling its approach:

1. Observable Assessments: Exams, vivas, and practicals will replace suspectible written submissions in critical courses.
2. Ethical Frameworks: Students will declare AI use in drafts, reflect on outputs, and engage in discipline-specific debates about AI’s role.
3. Curriculum Redesign: Lecturers are piloting AI tools for multilingual support and creative problem-solving through UCT’s AI Teaching Innovations Project.

"AI is ubiquitous in future workplaces," Walji argued. "Our graduates need critical literacies—not just avoidance—to use these tools ethically."

Lecturers & Students Respond: Relief and Uncertainty

Reactions to the policy shift are mixed:

• Faculty Concerns: Some staff worry about losing leverage over plagiarism.
• Student Relief: Many welcome the end of "policing" via unreliable detectors.

UCT’s open-access guidelines on AI integration aim to set a global precedent. "Universities must share strategies as AI evolves," Walji urged, citing collaborations with international peers.

The Bigger Picture: Can Academia Outpace AI?

UCT’s gamble hinges on a radical premise: If you can’t beat AI, teach it. By prioritizing assessment innovation over detection, the university bets on cultivating ethically savvy graduates—equipped to harness, not hide, AI’s potential.

"This isn’t surrender," Walji insisted. "It’s about redesigning learning for the AI era."