Who remembers SOHCAHTOA, BODMAS and all those other acronyms that helped us survive difficult formulae and concepts? The objective was to remember, recall and reproduce.
My undergraduate years needed a fair amount of cramming. I remember surviving Chemistry exams largely through determination and panic because MMUST insisted that I study Chemistry in first year despite having dropped it in Form Two.
Back then, knowledge was mainly found in lectures and library books. Accessing information required you to either sit in the library or borrow books from the long-loan or short-loan section. Even copying required effort. You first had to read something before reproducing it. Term papers were handwritten. I still remember one particularly creative friend who tore pages from a book and pasted them directly onto his assignment because he had run out of time to copy.
By the time I got to postgraduate studies, things had changed. Exams and assignments had moved to digital platforms. Plagiarism became easier because copying and pasting was now a possibility. One could borrow heavily from journals, rearrange a few sentences and confidently submit what appeared to be original thought. Supervisors and examiners, however, remained fiercely committed to reference lists. Sometimes it felt as though the bibliography carried more weight than the argument itself. Woe unto you if you did not know software like Zotero.
Then came Artificial Intelligence.
I would love to attribute my failure to pursue a PhD to the rise of AI. It sounds much more acceptable than blaming “climate change, government policies, and lack of infrastructure.” But for sure, as AI continued to rise, I thought that something significant would have to change in education.
The rigour of postgraduate study has traditionally been found in reviewing previous work. Over the years, accessing journals and research became easier, but one still had to read. At the very least, you needed to go through abstracts, determine relevance, extract useful ideas and decide how they fitted into your own argument.
Today, AI can find the papers, summarise them, organise the ideas and even help rewrite those ideas into a format suitable for your work. Reference lists no longer require an entire afternoon of frustration. You only need to know whether you are using APA 7 or another style, and the machine handles the rest.
What exactly should postgraduate education be examining now?
Communication scholar Everett Rogers argued through the Diffusion of Innovations Theory that institutions often adopt new technologies much more slowly than individuals. Looking at universities today, I wonder whether we are witnessing exactly that. Learners may have already integrated AI into their daily academic work, yet many assessment models continue to operate as though the technology does not exist. The result is a growing mismatch between how knowledge is produced and how knowledge is examined.
If research, particularly literature review, was previously valuable because of the effort required to gather, sort and synthesise information, what happens when a machine can perform much of that work in minutes? I struggle to see the value of assignments that merely require recalling content. Recall is probably the weakest way to assess a learner who has access to AI.
Do we return to oral examinations? Do we assess innovations and practical problem-solving? Do we ask learners to defend their ideas in real time? How do we make education meaningful in a world where information is available instantly?
At the moment, there are situations where learners use machines to generate answers, lecturers use machines to check answers, and other machines help mark them. Human beings are increasingly becoming supervisors of conversations between machines. Somewhere in that process, actual learning risks becoming optional.
Perhaps the bigger question is whether it is time to rethink education itself.
For decades, examinations have rewarded the ability to store and retrieve information. AI has dramatically reduced the value of that skill because information retrieval is now available on demand. Continuing to examine learners primarily on recall is like testing someone’s ability to perform long division while they are holding a calculator and pretending the calculator does not exist.
If machines are increasingly better at storing, retrieving and organising information, should human assessment continue rewarding those same skills? Modern learning theories such as constructivism suggest that education is not simply about acquiring information but about creating meaning, solving problems and applying knowledge in context. If that is true, then assessment may need to move beyond recall and toward demonstrating understanding, judgement and originality.
My view is that, for example, literature reviews should carry very little weight and perhaps reference lists should carry none at all. The machine can do those parts efficiently. Let it.
Instead, let us identify where AI should do the heavy lifting and make that official. Then let us identify where we genuinely need human judgement, creativity, ethics, problem-solving, curiosity and critical thinking. Those are the areas where brain cells should still earn marks.
No human being will outperform AI at tasks designed for AI. Equally, AI cannot fully replace the uniquely human ability to make sense of messy realities, exercise judgement and create meaning from experience.
Perhaps the future of education is not deciding whether AI belongs in the classroom. That debate is already over. The real question is whether our assessment systems can evolve quickly enough to distinguish between what machines should do and what learners should still be expected to do. If AI is now part of the knowledge creation process, then education must focus less on rewarding retrieval and more on rewarding insight. That, in my view, is where education should be headed.
References
- Diffusion of Innovations — Rogers, E. M. (2003). Diffusion of Innovations (5th ed.). New York: Free Press.
- Diffusion of Innovations — Rogers, E. M. (1995, 2003 editions), widely used in communication, technology adoption and organizational change studies.