Writing in the Age of Artificial Intelligence

AI is a widespread conundrum facing the literary world. Public debate has recently been triggered by Hachette’s pulling of the novel Shy Girl, questions about the originality of a 2026 Commonwealth Short Story Prize regional winner and the practices of Nobel Prize-winning author Olga Tokarczuk. As controversies continue to pile up, it seems that even the biggest players in publishing are woefully unprepared. We asked working writers about what the future looks like to them, their experiments with large language models and how they’d like to see the industry respond to this potential existential threat to their livelihoods.

Yassmin Abdel-Magied – At Sea

I can be a bit of a pedant about the word, or the acronym, ‘AI’, both as a craftsperson of language and as an engineer. Let’s be precise: when we casually refer to AI, I suppose what we might mean is the use of chatbots powered by large language models (LLMs, not exactly the same as ‘artificial intelligence’, especially given the concept of intelligence itself is contested). Am I against the use of this particular type of technology in the development of one’s craft? Not necessarily. However, I do believe there are certain considerations—ethical, critical—that mean I am distinctly against utilising the product options most of us have access to. The majority of the models available are built from stolen data; I prefer not to put money in the pockets of criminal enterprises when I can help it. The data inputted into most chatbots is often used to further train and improve models; I prefer my labour not to go towards lining the pockets of soulless billionaires, etc. There are some LLMs that are open source, built on non-copyrighted data, and that do not store or use your data, and so I am not wholesale against the use of the technology itself. However, by and large, I cannot help but think: we’ve been here before, people. Pay attention!

I don’t think the proliferation of generative language models has changed my approach to writing. What I write about tends to explore the minutiae and specificity of a particular type of human experience, one the models tend to not be trained on. I have done experiments with it—well, the engineer in me could not resist! As such, keeping in mind the concerns outlined earlier, I have tried a few open-source models, run a few locally from my machine and even run a few of my own scripts through the platforms to see if they would produce useful insights. I tend not to use any owned by the majors—Google, OpenAI, Anthropic etc, do not need my support, but Tinfoil, Lumo, Ollama and Apertus are welcome to it. I have found the models can be useful at the level of proofreading and structural analysis (notes on pacing, on whether all the characters are fully developed, etc), almost as a sounding board in lieu of harassing my editors and friends on ‘what they think of this new version of the text’. I have also found value in using the models for searching through large documents, or as a first-pass literature review when delving into researching a new topic, helping me unearth references I might not have found for some time—or ever. But for idea or prose generation? I have found on those fronts these models are a waste of time. I mean, that’s not what they are designed for, and plus—that’s the fun part of this job! I wouldn’t give that away for love nor money. What else is the point!

Ed Coper – Angertainment

The temptation is real, because I’ve never had to use my brain’s ‘compute’ power more than when writing a book—and AI gives humans compute power that is greater than our own. But I don’t because my goal is not to be a supercomputer but to connect with real live humans.

In my darker moments I wonder whether creative writing is dead—to both the hand, as writers outsource one of the most difficult human endeavours, and to the ear, as readers lose the ability to connect with ambiguity and novelty when surrounded by texts that are all the same. But then I remember that AI is essentially just a mirror. It can only tell us what is already known, in prose that mimics what is already said. And that all my favourite creative works (from books to music to films) are unique because they did things nobody had ever done before. And only humans are capable of extending what humans are capable of. Let creative writing be the last stand against AI-replacement, as the most human of humanity’s feats. Let it eliminate all the meaningless ‘busy work’ of our lives so we all have more time to read.

Olivia De Zilva – Sick Note 

Using AI in a creative practice takes away from the unhinged nature of the human voice, which is what makes stories relatable and intimately felt by readers. I feel like creativity needs to be earnt and toiled over rather than given an easy way out. That’s the fun/torment of it, right? In the context of a corporate job, it can be helpful in organising things and creating workflows, but there is a perception that it can support workers to be more efficient until we are not needed at all. This is something I struggle with as we are not robots, nor do we have expendable and endless energy for work and capitalism. I’ve also tried to use it to plan a holiday—it was rubbish.

Emily Gale – The Wild Unknown

Once upon a time I’d hoped to ignore AI for the rest of my life, but somehow I ended up writing a children’s book about it. My research confirmed that I should never go near it again. I wish it was as simple as saying that AI is pure crap, but the situation is worse than that—AI is quite good and it’s much easier than doing it yourself, which makes it the enemy of creativity. Although the image of a lone, struggling writer surrounded by crumpled paper balls is a cliché, aloneness and struggle are at the heart of our creative drive, our deep understanding of what we’re creating and the euphoria we receive at the end. When we see struggle as suffering, it’s too tempting to turn to external problem-solving—in come the AI apps calling themselves ‘a thinking partner’, ‘a novel crafter’, ‘the best model for creative ideation’. With their negative patterns (sycophancy, deception, hallucination, etc), and our negative patterns (self-doubt, ego, poor self-discipline, etc), I believe it will be a short road to a state of dependency, and with that comes the slow but sure rotting of the creative muscle.

Anna Goldsworthy – Quarterly Essay 102: The God We Made

The genie is out of the bottle. How much it is deployed will depend on the tolerance of writers and of readers. Many writers already rely on algorithmic enhancements—from spellchecker to Google Search to the interventions of Grammarly—so the argument is one of degree rather than any sort of moral binary. And ChatGPT can be used in many ways: as research assistant or interlocutor or structural aid or line editor or outright text generator. I picked up a book at the airport today which breathlessly announced it had been produced largely by ChatGPT, and that the ‘writing’, such as it was, had consisted of editing and revision—allowing the writer to produce four books in the time it would previously have taken her to produce one.

But I wonder what we give away as we abandon the slow work of our brains. One of these losses might be voice. ChatGPT’s own style has been widely documented—the telltale vocabulary, the symmetrical sentence constructions, the rule of three. But studies have proven this can be overridden by training in a specific author’s style. And so the ‘voice’ becomes the unique object rather than the text—and perhaps the locus of copyright. Further research shows that ChatGPT’s house style is now permeating the output of academics, even when speaking extemporaneously, drawing human expression back to the mean as it reduces linguistic biodiversity. How does a young writer hatch a style of their own if constantly subject to the mediations of AI? Personally, I have a sentimental attachment to brain-made work (but then I am currently scribbling these notes into my notebook with a fountain pen, so perhaps a fetishist). Above all we risk sacrificing that large part of the literary experience that is relational: the intimacy of an encounter with another human mind.

Allanah Hunt – Forever & Ever

I feel AI doesn’t have a place in our creative world. Using AI to write is not a victimless act; AI does not pull its knowledge from random sequences of numbers. It uses knowledge it coalesces from other authors, other authors who have worked tirelessly to improve their craft and convey an idea to a reader in the most engaging way possible. Your individual voice is something you only find after months or years of practice. No one else can recreate this; they can only copy but it will never be theirs. An AI-generated text has no individuality, no soul and it will never become one-of-a-kind. It is all about mimicry, not creation. Other people should not unfairly benefit from another writer’s creative prowess. I see AI as an extractive tool, mining for resources and leaving the Country behind bereft and barren.

I worry that AI will limit our imagination as our brains start to only stretch to the capacity of what an AI can comprehend, not our own complex understandings of the world and ourselves. I’ve seen AI be used in place of critical thought and I fear it will be used in place of creative thought as well. I would hope publishers take a hard line that AI ‘creative’ work doesn’t have a place in the publishing world. I think it is losing the soul of creativity if we no longer value a person’s thoughts and heart but instead turn to an extractive reworking of the same idea, over and over again, until there is no originality or sparks of its origins left.

Sheree Joseph – Juicy

Bad writing is so prolific that the reason we couldn’t differentiate between AI and everything else is that we just thought it was the same bad writing we’ve always seen. So, to me, this is really more a battle between bad writers, who have always existed, and bad writers with AI, and that doesn’t really interest me. I am interested in being a good writer, and good writers are always questioning themselves, always thinking they could do better, always obsessing over every word, forever editing, forever thinking up new ways of saying things, taking risks that don’t pay off and never feeling fully satisfied. Good writers would never be mired in this debate because the very idea of AI would make their stomachs churn—this thing that cannot be original, that cannot take risks, or understand what it feels like to be human in that very strange and weird way that even humans struggle to understand and spend lifetimes trying to capture in language.

Having said this, I am increasingly uncomfortable with the AI discourse and the sudden witch hunts. When will we learn that going after individuals doesn’t make sense when entire industries are pushing this shit and making major redundancies across sectors in favour of forced AI usage. We’ve seen the long-running battle well before this—the ditching of subeditors and the way skilled labour has been deprioritised. Who bears the brunt of this? Writers. Always writers. As an AI hater from day one, I get the frustration, but it’s just so bizarre how the discourse has turned us against each other, and in ways that are wholly dubious. (Like ‘AI detector tools’ that are taking writing from pre-AI eras and saying it’s 98 per cent likely to be AI because of the highly polished, grammatically flawless and ‘lack of human inconsistency’ prose—failing to understand that writers work hard to get it as right as they can.) This obsession with individuals masks the true problem, like putting a Band-Aid on a gushing wound. Start going after systems—those at the top making terribly misguided decisions.

Isabelle Li – The Northern Tomb

A friend of mine is a nerd. He was the first to tell me about generative AI, in one of his customary unannounced afternoon visits. I switched on my laptop and was lucky to be granted access to ChatGPT immediately. We asked the chatbot to write limericks, tell jokes and produce comic sketches by applying incongruity theory. Some worked better than others. We had a good laugh over a pot of osmanthus tea.

As a translator between Chinese and English, I’d been tempted to test out Google Translate and ended up having to rewrite and reshape, much like starting with a page already covered with unsatisfactory drawings which I had to erase before putting down my own lines, doubling the work. A few paragraphs were enough to convince me such ‘convenience’ was but a trap. It’s however a useful tool for sanity checks. I once reviewed a lovely poetry translation manuscript with an entire stanza missing. Google Translate is a quick way to test for completeness. As a late adopter, I’ve decided to embrace AI and find it handy for preliminary research, transcribing Teams meetings and drafting minutes. I wouldn’t use it to write though, for the same reasons I wouldn’t use Google Translate for literary translation. I liken AI development to gain-of-function research, a precarious play that poses existential risks, which must be regulated nationally and internationally. I also hope that governments around the world will foster competition to break the concentration of AI power. I support publishers/publications imposing strict policies to exclude AI-generated content as plagiarism should never be permissible. However, there may be a need to establish a separate prize category, herding bots to a different lane from the human marathon runners, if their race is of interest to spectators.

Today AI-generated content permeates public space. We breathe in a plastic language, mindless text based on probability. Large language models mirror our brain, a powerful predictive engine. But with the advantage of consciousness and embodied experience we can interrogate the language we use, sharpen our thoughts and expressions, revise the text and occasionally transcend our limits and discover something new. The recent case of a major prize awarding seemingly AI-generated stories ignites a worse fear. What if the writers weren’t lying? What if their own writing is indistinguishable from that of AI? It would be a watershed moment in our literary history, the start of a fall towards the black hole of homogeneity from which no one can escape.

Wayne Marshall – Henry Goes Bush

In a truly dark time for the future of the human imagination, the one shining light I offer is this—the robots haven’t got shit on us when it comes to the weird, the unclassifiable and the idiosyncratic (ie, all the best stuff). Sure, AI seems frighteningly adept at absorbing and spewing back at us texts that might fall under the banner of ‘formulaic’ and ‘predictable’. Equally frightening are the writers unfazed by boarding that inglorious train. (Where’s the hard-won satisfaction in it? Where’s the fun? What of all those terrible daily doubts and fears about the work that over 4, 5, 6, 7+ years can be transformed into something golden?) But the speck of good news is that the machines (as yet) haven’t found a way of capturing the strange corners of our imaginations. They can’t imitate the voice or style of a writer whose work roars onto the page—they don’t have access to those mystical frequencies. So yeah, things aren’t great. But if all this forces a higher value on work that gallops in its own direction, on storytelling that thrives beyond the reach of AI’s slimy tentacles, I consider that a silver lining for sure.  

Steve MinOn – First Name Second Name

I wish I’d known about AI before I wrote my novel. The one that explores complex feelings about my identity. It might have been quicker. As easy as ordering a pizza online—one half ham and pineapple, the other half crispy skin chicken with lup chong, bok choy and hoisin sauce. And while AI wrote my novel, the one with a half-Caucasian, half-Asian voice, I could have been making myself a cocktail of Italian Vermouth, Chinese Baijiu and Tonic Water to drink with it, while my mixed-race, culturally lapsed, gay self-loathing, historical fiction and body horror novel wrote itself. It would have been so easy.

Prasanthi Ram – Nine Yard Sarees 

I write stories because I am curious; I have questions that intrigue me but don’t know how to answer. So writing becomes a kind of wayfinding. I wade happily through my own mind, my experiences, my observations of the world. Losing my way in that process is also very normal and, perhaps to the surprise of some, what truly invigorates me. On the flipside, AI, with all its mathematical optimisation, is a Great Bypass. You don’t even get a chance to be lost. To be uncomfortable, to be bothered, to realise that perhaps the question you began with isn’t quite the question you wanted to ask. So this is my long way of saying that AI hasn’t changed my approach to writing fiction at all; after all, petty as it sounds, I came before it, and so did my writing. If anything, I write against what AI signifies in our world today: a gross flattening of what it means to be human.

Allee Richards – Tight Lines

Mid last year I entered a transcript of an interview I’d conducted into an AI engine. I asked the bot to generate a feature article from the contents of the interview. I gave it a pitch and included links to several articles I’d previously published as examples of style. What came back was absolute drivel. Relieved, I went back to work, no longer fearing my part in its future. What I didn’t realise then was that the creep of AI would not be in realising entire articles but in finessing sentences. Reaching the end of a paragraph and realising I have used the word realise four times, one can ask the bot: can you reword this for me?

The theatre director Beng Oh once described a play to me as ‘one part of the ongoing conversation of what it means to be alive’. I don’t believe AI has a place in any form of human expression. And yet the temptation to take the easier, less rewarding and faster route is always there, niggling in the background like the desire to check Instagram or my emails when I should be enjoying any given moment.

Ellena Savage – The Ruiners

It’s clear why an epoch of lifestyle disease produced an obsession with training and conditioning the body. The same forces are at work towards mutually constitutive ends. When intelligence is more or less completely outsourced to algorithms at the population level, I imagine there’ll be a similar mania for training and conditioning the intellect. The problem with this is that thinking will become a hobby for people with time and money, an elite hobby, a brand of self-discipline, rather than a basic right and a precondition for full participation in society. Personally, I’m anxious that I’m getting slower and lazier every day, that everyone is getting slower and lazier every day, and that I have to put myself in a state of frantic activity and opposition to retain the right to use my mind. As a writer I would like to direct this energy towards writing the best work I can. I don’t know if that’s working. This is starting to feel like tap dancing while shark-infested hypothermic water enters the hull. Are we just archiving human culture before it’s all gone forever? Are we just feeding the algorithm to enrich a class of psychopaths? I don’t know.

After the Commonwealth Story Prize scandal, my brother, a STEM person, made the point that the ‘humanities crowd’ doesn’t spend enough time with AI-generated text to recognise its tics, and this makes us look naïve. I understand when I’m being razzed, but he’s right that our sensitivities can inure us to reality, and it’s not good enough. Saying that, there’s also a problem in the literary space where a kind of impersonation of ‘literature’ is granted the status of literature. So what sounds ‘literary’ (ornate, breathy, motif-heavy prose) is repeated and reified as in fact literary, when it’s dull, repeatable and embarrassing. The impersonation could be achieved by a human being or an artificial intelligence. Voice is particular. You can’t fake it. It’s to do with having a personality, which AI doesn’t, and which some people in the literary business don’t either. I’d like for publishers to not collaborate in the certain destruction of their businesses and the annihilation of human culture. And if you’re interested in outsourcing your cognitive activity to word-prediction algorithms designed by the world’s most diabolical losers, don’t be a writer. Just do anything else.

I’d like to say I’ve never used ChatGPT, but I did use it once when I was heartbroken. Why is this happening to me? I asked it. Why doesn’t he love me? The word predictor gave me the therapeutic platitudes it thought I wanted to hear, and this confirmed that I’d lost control of my mind. Heartbreak is literally humiliating; you’re diminished by the temporary withdrawal of intellect, reason, self respect, moral dignity. It’s reasonable in this condition to seek counsel that confirms that you’re good, actually, and the other person is clinically insane. But even when you’re suffering indignity, crouched in worm position, reality is still somewhere in the room. AI won’t tell you this, but life is unfair and most people are crazy. Luckily, life is unfair and most people are crazy is the greatest topic, and the true subject of literature.

Ilka Tampke – How to Love the World

Being a novelist in the age of AI has taught me what actually matters about creative writing. Reading the kind of soulless fiction produced by AI (I teach creative writing at university, so I read a lot of it) has helped me understand why anyone would still labour for days, weeks, years to birth their ideas into language, when they could press a key and do it in seconds. Voice and style in writing are everything. Grammar is how we sculpt the clay of language and the way we use it—the idiosyncratic quirks of our syntax, punctuation and vocabulary—is entirely individual, as unique as our face, our speaking voice, the odour of our skin. No two people see the world in the same way, nor express their view with the same metaphoric lexicon. When you strip this from writing, you are left with something that looks and acts human but somehow isn’t. We know this figure—the Zombie. For me, writing a novel is very difficult. It takes me to the limits of my capacity, discipline and faith. Just as I wouldn’t watch an Olympic marathon runner who used an AI bot to run their race, I wouldn’t read a book that a human hadn’t toiled to create. It is the struggle, the yearning and the courage embedded in a novel that make it mean something. To read it is to dwell, for a time, inside another human consciousness. This is why it activates our brain’s mirror neuron systems, which undergird our capacity for empathy. We read to connect to each other.

There is already superintelligence in the world that is far greater than us, to which we can turn to help us write. It is not artificial—it is ecological, or perhaps (given plants make up 80 per cent of the world’s biomass) it is vegetal. I have been lucky enough to spend the last thirteen years in a part of Victoria where I can regularly walk among forests. Trees do not have brains or words, but they do speak in an eloquent language of chemicals, shape, colour, scent and growth that contemporary science is only just beginning to understand. When I am immersed in a plant world, ideas come quickly, I think more freely, more associatively. Creativity does not belong to humans; it is the fundamental nature of planetary life. Ironically, it is this genuine, ecological intelligence that artificial intelligence is imperilling. In an age of extinction where we are facing catastrophic climate change and biodiversity loss, we are ushering in a technology that uses water and power at a scale unprecedented in digital history. I used AI once. I gave it a brief synopsis of a book I was writing and asked it to think of a title. Its suggestions were clichéd, but it complimented my synopsis and, tipsy on dopamine, I felt an all-too-human urge to prompt it for further affirmation. That’s when I shut my laptop and went for a walk. AI will almost certainly come to produce creative writing that appears less soulless. But that doesn’t mean that it won’t still be soulless. Whether that matters is a question for each of us to reckon with.