Much as it pains me to admit this, it took me over a decade to learn how to properly teach the humanities in South African universities. In a way, I have ChatGPT, the artificial intelligence (AI) chatbot launched by Open AI research laboratory late last year, to thank for this.

Long before news of the arrival of ChatGPT sparked panic in the richly endowed universities of the global North, southern African university lecturers were faced with the pervasive use by students of bottom-feeder, synonym-swapping web tools in their assignments.

This was true before the Covid-19 pandemic intensified the move to remote, online assessment. Students would often submit responses to prescribed readings and essays which attempted to disguise plagiarism by pasting chunks of prescribed text into free-to-use websites, and with the click of a button, synonyms are swapped, often to unintended comic effect.

In one instance that has become the stuff of legend among local scholars, a student submitted an essay where a synonym swapper transformed Fort Hare University into “Stronghold Bunny College”.

Plagiarism is as old as Gutenberg’s printing press. Like generations before them, contemporary students plagiarise for perfectly explicable reasons: a lack of confidence in their comprehension of prescribed readings and in their writing (often not in their first language); because they’ve left their work to the last minute; and, paradoxically, because they are constantly being warned against plagiarising and must already feed their work into plagiarism detecting software.

Or, most alarmingly, because they don’t fully understand what plagiarism is, and why it is a problem. Many of our schools teach students to regurgitate “textbook answers”, or what teachers insist will score best in exams.

Since ChatGPT’s release, university lecturers across the globe have been wringing their hands and strategising about how to address the problem that ChatGPT can — with varying degrees of sophistication — generate answers to questions in ways that are more difficult to detect than old-fashioned forms of plagiarism. Some of the world’s elite universities have already acquired AI technology that detects whether ChatGPT has been used.

Locally, some of my colleagues are arguing that we can ride this dragon: ChatGPT can be put to creative use, they say, and even improve the writing of students. While I respect this attempt at avoiding knee-jerk Luddism, these arguments are unconvincing, even foolhardy, in an age where democracy itself has been thrown into a global crisis partly because of the increasingly deleterious effects of the internet on public debate.

Happily, ChatGPT forced my hand. Already fed up with marking work littered with poorly selected synonyms, or simply copied and pasted from prescribed texts, for my African History Honours seminar this semester, I decided that the time had come to try a radically old-fashioned approach.

As usual, students are given a prescribed reading each week, and upon arrival at class, a double-sided blank sheet of paper containing a question they have not seen beforehand, to which they are required to respond, without using their notes, or having the prescribed text at hand.

Ironically, this method exploits the seriousness with which students treat examinations.

Here’s the thing: not only am I no longer marking dispiriting synonym soups, but I’ve never had such high-quality in-class discussions. This, after years of naïvely hoping that students will complete, comprehend and freely discuss readings in the mode of a humanities seminar, and only finding that a handful of students do so.

The real crisis we face in the age of bots is not writing, but the endangered skill of discerning reading, which fuels better thinking and writing. We need to defend and deepen this skill before it is lost to the majority of us, forever. DM