To AI or not to AI? How chatbots can help revive the university essay

Generative AI does not signal the end of student writing. Even that time-honoured (if shop-soiled) form of writing – the “student essay” – might be on the verge of a revival. But only if we take the time to reconsider the history and the real purpose of essay writing.

Essays might never look quite like they have done for the past century or so but, honestly, who would really mourn the passing of the kinds of formulaic thinking that often came with them?

Artificial intelligence can help uni students write essays. But is that such a bad thing?Credit: Getty Images

Why, though, do we get so anxious about uni students not doing their own writing in the age of ubiquitous AI? It is because an essay is not just an indication of whether somebody knows something. Rather, we value the task of writing essays as a way to develop crucial skills in evaluating knowledge and communicating ideas.

From the first notes jotted down in response to a piece of evidence, through analysing other written sources, and then onwards to drafting a plan, and eventually to editing, revising, and finalising a piece of work, the time that all of this takes aligns with the capacity of our brains to dream up ideas, to evaluate them and then try to express them.

Artificial intelligence squashes that time into seconds. Each of these stages can now be done almost instantaneously. In an education system where we give grades to the end-product of the “essay” rather than to the process of thinking and writing, and where students’ time is increasingly limited by part-time work and by busy curricula, then of course the instantaneous has its attractions.

Essays were not always like this. My own area of academic expertise is in 16th- and 17th-century literature, the period in which the very first “essays” were written. The first writer to use the term to denote his work was a French lawyer, Michel de Montaigne (1533-1592). Over the course of his life, he wrote 107 pieces of work that he collected under the term, “essais”.

The French verb, essayer, means to try, to attempt, to test, to take the measure of something, to work out what something is worth. Montaigne’s writing is brilliantly various and idiosyncratic, taking in topics from cannibalism to friendship, from war horses to sadness, from solitude to drunkenness. And what we get in his work is a kind of shuttling between individual experience and shared knowledge: how does my experience of this thing measure up to what other people have thought about it over time?

Such writing is often partial, idiosyncratic and individual. When he discusses his own knowledge work, Montaigne thinks about it as an ongoing process that unfolds across time – unfinished, circumstantial, self-consciously limited by both his own experience and the material upon which he is working. But it is also a kind of thinking and writing that is willing to meet with new circumstances, and with no specious pretence to universal knowledge and understanding.

My hope is that a reinvigorated sense of what an essay might do would get back in touch with this version of the “essay”: one that is less pressured to be “right” (pressure that leads to short cuts) and is more concerned with testing individual understanding of a topic against shared knowledge: changing and ephemeral. These “essays” would be exploratory rather than definitive, contingent on developing student experience rather than universalising.

How could this work in practice?

In my own institution, The University of Sydney, we have been engaged in a project to rethink student writing. Breaking away from the formulae of the past, we are driving the practice of writing-as-thinking back into the classroom.

In my own “Shakespeare” unit, students still write essays about Shakespeare but many of their ideas are generated in class, in real time, and in response to evidence that they only encounter for the first time in that setting. They can then go away and re-evaluate what they have written, but this practice ensures that individuality and idiosyncrasy is now at the heart of the work. And, as a teacher, I get to show students how much more I value the “process” than the “product”.

And critical use of AI can also get a look in: a vital skill to develop in our students. One of the concepts that I have noticed my students struggling with is spotting whether Shakespeare is using a word metaphorically or not. In response to this, I have been developing a chatbot using the university’s AI platform that enables them to work it out. It doesn’t give them an easy yes/no answer but engages with their examples in a way that develops their understanding of how his language works.

And, like many of my colleagues, I help students to see how AI tools could be used to generate ideas (identify sources; summarise key concepts) rather than merely work as a tool to imitate the work that we want them to do.

The answer to the problem of the student essay in the age of AI is to lean into its original motivations. Like Montaigne, we want to see students thinking through their individual ideas, testing them against shared knowledge.

And it is also to slow our students down. If we are concerned about the speedy off-ramps offered by generative AI, then we break the “essay” down into something that is more fragmentary, that takes place in deliberate stages and doesn’t always try to rush to cheap conclusions.

Associate Professor Huw Griffiths teaches and researches Shakespeare’s work at The University of Sydney.

The Opinion newsletter is a weekly wrap of views that will challenge, champion and inform your own. Sign up here.