Is there any point still teaching academic writing in the AI age?

University students, virtually everyone agrees, ought to learn to write well. Writing is an essential skill for academic and professional life, and so, across disciplines, we teach students to compose essays, reports, research papers and literature reviews. We want them to develop a command of language, an understanding of structure and form, and maybe even an original voice. However, in the age of AI, what it means “to write” is changing.

Those of us above a certain age might still picture a person sitting at a desk, or in a coffee shop, surrounded by books as they tap out a first essay draft on a keyboard. But in the not-too-distant future, a student given a writing assignment will be more likely to start by dictating her scattered thoughts and half-baked ideas into a phone app that transcribes them into text. Next, she might feed those dictated notes into a writing assistant such as ChatGPT, asking it to convert her observations and insights into a coherent essay or forum post. Now she has a rough draft. Perhaps she will then sit down with her laptop to thoroughly edit it – or perhaps she’ll just make a few quick changes on her phone before submitting. 

For those of us who teach writing – those of us who spend our days preaching the importance of outlines and proofreading – it’s important to think about how the processes of writing are changing, because our teaching will have to change accordingly. 

To my mind, there’s not much point being romantic about the bygone days of robot-free writing: the terrors and possibilities evoked by staring at a blank page, not knowing where to begin. Cursive used to be an essential skill for writing quickly by hand, but who even carries a pen now? Eventually, primary schools will stop teaching it. Perhaps, as students get used to dictating their thoughts to a computer, typing, too, will cease to be a relevant skill.

However, while change might be inevitable, we should not greet it with indifference. Technological change is normal; it is never neutral. As well as making familiar habits and ways of life impractical or impossible, they place new demands on us. They necessitate new policies. If AI-assisted writing becomes the new normal, teachers will have to reassess the purpose and practice of writing education.

Many commentators have suggested that traditional teaching strategies will be rendered obsolete as it becomes impossible to distinguish between authentic student writing and that farmed out to large-language models, such as ChatGPT. Personally, I have my doubts that that day will ever arrive. After all, Silicon Valley is littered with failed companies that overpromised and underdelivered. AI writing assistants are still deeply flawed tools, and they might never be able to match the creativity, sophistication and subtlety of human writers. Still, when competent writing can be produced with the push of a button, persuading students they need to develop those competencies themselves certainly becomes more challenging.

Prudence dictates some pedagogical adjustments. If AI writing assistants make it easy to produce drafts, the actual work of writing will shift to the editing process. Rather than striving to meet word counts by turning their thoughts into sentences, young writers will be more like sculptors, trying to craft something meaningful and precise from the rough block of generic text that AI has provided them. We already teach students about proofreading and revising their work, but perhaps this should ultimately become the whole focus of our writing pedagogy.

The rhythms of writing assignments – the essay, the research paper, the thesis – are all deeply integrated into the very structure of higher education. If writing assignments become unreliable tools of assessment, many aspects of higher education might need to be transformed.

One thought is that if writing becomes easier to produce, professors could demand more of it – though, clearly, assigning 15-page papers instead of five-page papers would make for an untenable amount of marking if done routinely, so we’d have to choose our moments. Alternatively, should we simply expect better writing if every student has an AI writing assistant at their disposal?

Or dare we to leave writing in the dust entirely? Perhaps we “pivot to video”, as most major media companies on the internet have done. Maybe creative assignments or class presentations could replace some traditional written work. What about oral exams? These have their place, though logistics are an issue. Can you fit 60 student interviews into your schedule?

If we stick with written assessment, our main objective might now be to resist the great stylistic flattening sure to ensue as AI writing becomes more common, working with our students to ensure their individual thoughts and perspectives are not lost in a sea of bland, easily digestible copy.

Arguably, a banality of written expression has long been a problem among student writers – and, if we are honest with ourselves, we can admit that our own stale assignments have contributed to the problem. But now is our opportunity to up our game.

In adapting our teaching to address the challenges posed by AI, perhaps there are opportunities for us to recommit to developing excellence, originality and even style in student writing.

Andrew Moore is associate professor in the great books programme at St Thomas University in Fredericton, New Brunswick.