Picture: Canva
Gus Silber writes elsewhere on the internet:
A seismic shift shook the Corinthian and Ionic pillars of South African academia this week, with the announcement by the University of Cape Town that it was abandoning the use of AI detection software for academic assessment.
No longer will discredited tools such as Turnitin’s AI content checker be deployed, like the judges of Salem, to divine whether a student’s work was shaped by a magical incantation.
Or as we call it nowadays, a prompt.
It would be all too easy to interpret this move by UCT, the top-ranked university in Africa, as the waving of a white flag, a sigh of surrender to the ghost in the machine.
In truth, it is an about-time acknowledgement of an age-old reality, that when a disruptive technology seeps into our everyday lives, it can and will no longer be wished away.
I remember the furious debate, in my matric year, over whether learners should be allowed to use handheld calculators during maths exams.
Authoritarians of the Old School argued that the devices would erode memory and mental acuity, in much the same way as Plato once argued against the technology of writing: "If men learn this," he declared, "it will implant forgetfulness in their souls."
Eventually the New School won the day, and we were allowed to tap, tap, tap away as we toiled over our equations. Not that it helped me much, alas; I still got a G for maths.
Likewise, when I worked on the Sunday Times in the early 1980s, I had to give up my beloved Olivetti manual typewriter for a computer terminal on which words flickered in luminous green type.
The clattering of mechanical keys in the newsroom soon gave way to the bellowing of reporters whose stories had been swallowed whole by the monster in the machine.
But technology learns from its mistakes, just as we learn from technology, when and where to use it, and when not.
I recently took part in a Jive Media webinar on AI in academia, and one of my fellow panelists, Michelle Riedlinger, an associate professor at the Queensland University of Technology in Brisbane, came up with a lovely metaphor to highlight the difference between machine thinking and human thinking.
She spoke of "forklifting tasks" versus "weightlifting tasks", the former being repetitive, time-consuming tasks where you can let the machine do the heavy lifting for you.
In the academic context, that might mean transcribing interviews, translating research, and mining voluminous PDFs in search of nuggets of data.
But you're still the expert driver, you're still at the controls, and now and then you need to get out of the forklift and move things around by hand. That's where the weightlifting comes in.
"Weightlifting tasks are where we want to build strength," said Michelle. "Things like framing stories for particular audiences, learning about your audiences, and interpreting data for particular audiences."
If we offload too much of that weightlifting to AI, she added, "we risk weakening our long-term capability as communicators."
But let's move the forklift over to UCT for a moment.
The academy's decision to step away from AI detection tools, which are notoriously unreliable and prone to false positives, is part of a much broader strategy called the AI in Education Framework.
What this means, according to an article in the Sunday Times, reposted on UCT's website, is a variety of assessment methods, including oral exams, observed group work, and assignments in which students must disclose and critically reflect on their use of AI.
The goal is not simply to react to AI disruption, but to build a “compass and road map”, said Sukaina Walji, Director of UCT’s Centre for Innovation in Learning and Teaching.
Wherever that journey may lead, the signposts are clear: learning and teaching, in the age of AI, will never be the same.
When ChatGPT can spit out a passing-grade discursive essay in seconds – even if you have to add a "please humanise this for me" prompt to its output – what does that tell us about the new generation of students?
That they are lazy? That they are willing victims of the calm tyranny of convenience? Or does it, perhaps, tell us that the essay format itself is overdue for reinvention?
The educators who argued against the calculator, the journalists who swore at the computer, had to concede, eventually, that resistance was futile. The machine was a tool, not a threat.
In my own much-delayed studies, the epiphany that most jolted me out of my complacency of thinking I knew stuff, was an instruction that a master's thesis need not be a model of original research.
Rather, the student should grapple with, interrogate, and build on the research that has gone before, aiming to see further, as Isaac Newton put it, by standing on the shoulders of giants.
But now here we are, standing not just on the shoulders of giants, but on the shoulders of machines that have been fed and trained on all the learnings of all the giants that have gone before.
AI can echo the voices of the past, but it cannot replace the human struggle to understand, to question, to challenge, to interpret, to find meaning amidst the noise and chaos.
That is the true work of learning, to be both curious and critical, as Sukaina Walji of UCT reminds us.
In the end, it’s not about exorcising the ghost in the machine.
It's about doing whatever it may take, to keep the human voice alive in the conversation.