21 May 2026

Beyond the AI clampdown

By College SA

By Stefan Botha, CEO of Optimi

Recent reports that South African universities are tightening controls around the unethical use of artificial intelligence in assignments are not surprising. Across the sector, institutions are responding to a growing reality: students now have easy access to tools that can generate text, answer questions, solve problems and imitate original work at speed. Universities are right to protect academic integrity. A qualification must still reflect a student’s own competence, judgment and effort.

But the bigger question is this: what should education do next?

If the response to AI begins and ends with detection, surveillance and punishment, South Africa will miss the real opportunity. The issue is not simply that students are using AI. The issue is whether they understand the difference between using AI as a support tool and using it as a substitute for thinking.

That distinction matters, because the world students are entering is not AI-free. In the workplace, artificial intelligence is already shaping how people write, research, analyse, communicate and solve problems. Employers are not looking for graduates who have merely been kept away from AI. They are looking for people who can use technology responsibly, question outputs critically, protect confidentiality, verify facts and still take ownership of the final result.

That is why the current debate should push the sector beyond enforcement and towards literacy. Students need clear rules, but they also need clear teaching. They need to know when AI use is acceptable, when it is not, how it should be disclosed, and why ethical use matters. They must learn that convenience cannot replace competence. If a student cannot explain the logic behind an answer, defend a recommendation, or apply knowledge in a real setting, then the tool has not educated them - it has only assisted them to submit something.

This is especially important in occupational and career-focused learning. In these environments, the purpose of education is not only to pass an assessment, but to prepare a person to perform in the real world. That means students must leave with practical capability, not just polished submissions. It also means providers must design learning in ways that test what matters most: applied understanding, problem-solving, communication, accountability and the ability to make sound decisions in context.

The answer, then, is not a simplistic choice between banning AI and embracing it uncritically. The better path is structured, transparent and intentional use. International guidance has already moved in that direction, emphasising a human-centred approach to generative AI in education, while South African higher education research has similarly pointed to both the benefits of AI for teaching and learning and the urgent need for policy frameworks that protect academic integrity.

For institutions, this means rethinking assessment. More oral assessments, workplace simulations, practical demonstrations, staged assignments, reflective submissions and supervised evaluations can all help distinguish authentic learning from automated output. It also means making disclosure normal. Students should not be left to guess where the line is. Every assessment should state whether AI is prohibited, permitted in limited ways, or required for a defined purpose.

For students, the message should be equally clear: AI can help you learn, but it cannot learn on your behalf. It can assist with structure, brainstorming, summarising and refinement, but it cannot replace your responsibility to understand the work, test the information and stand behind what you submit. Integrity is not an old-fashioned academic principle. It is a workplace principle. Employers need people they can trust, especially in an economy where decision-making, compliance and communication matter as much as technical ability.

For South Africa, this conversation also has a wider significance. We are under pressure to build a more employable, adaptable workforce in a changing economy. That requires education providers to produce graduates and working adults who are digitally capable, ethically grounded and ready to work with new technologies without becoming dependent on them. The future will not reward those who avoid AI altogether. It will reward those who know how to use it wisely.

The current clampdown should therefore be seen as a necessary first step, not the final one. Protecting standards matters. But the long-term goal must be larger than catching misconduct. It must be to build a culture in which students understand that technology is a tool, not a shortcut to credibility.

In the end, education still has to do what it has always done at its best: develop independent thinkers, capable practitioners and responsible citizens. AI changes the context, but it does not change that mission. If anything, it makes it more urgent.