![](https://dl6pgk4f88hky.cloudfront.net/2024/09/05/Z326_029-1038x778.jpg)
As with many other technological advances, I responded to the rise of artificial intelligence by insistently ignoring it, hoping it would go away before I was forced to understand its significance. Many decry sticking your head in the sand as a general approach to life, but I find that, more often than not, it bears out. Submerge your head for long enough, and half the time the issue really does go away of its own accord. Not so this time. A few years ago, friends of mine who write in more technical and commercial capacities began to discuss their fear of being replaced by AI. When they showed me examples of AI-authored content, the writing seemed so clunky and ham-fisted that I couldn’t imagine it superseding even the most inadequate copywriter.
Cut to this spring, when I, having moved to one of the most expensive cities on earth, was trying to find some extra writing gigs. After a few days of sending emails and speaking to recruiters, the only work I was offered was training AI programs how to write. I was reminded of one of those adverts for meat, in which a beaming pig in a chef’s hat fires up a delicious pan of sausages.
Forced to concede that AI was, in fact, being used in precisely the way my friends warned, I tried to comforted myself. Yes, AI could replace many kinds of writing, including some that I do myself. But it could surely never produce literary fiction.
This week, the non-profit organisation National Novel Writing Month (also known as NaNoWriMo), which encourages aspiring writers to produce a 50,000-word manuscript during the month of November, posted a statement insisting that it “does not explicitly support any specific approach to writing, nor does it explicitly condemn any approach, including the use of AI”. Not only does it not object to fiction writers use of AI, the organisation believes that “the categorical condemnation of Artificial Intelligence has classist and ableist undertones”.
Such an objection, of course, absurd. But it arises from the somewhat plausible contention that not every aspiring writer can afford to consult a human editor. On a functional level this is true – and one could argue that the escalation from spellcheck to an AI editing programme is not so steep. But if we are to consider more seriously what we wish literature and writing to be in our culture, the obvious point is that AI is only capable of editing (or changing, or generating) a writer’s work to be ostensibly technically more proficient. A great editor identifies the particular value of a manuscript and works to chisel away the extraneous, making it not only better but more itself. AI, by definition, use pre-existing work to identify generic rules and formula, which it can project on to the work before it. To make something “better”, means to make it more like everything else.
Asserting an individual’s right to use AI to create a piece and still be considered the author leads to the erosion of the writing process. Believe me, I know better than anyone the desire to have written something, and the simultaneous repulsion at the thought of sitting down and writing it. All my life I dreamed of the day when I could write for a living – it was my greatest fantasy. And yet every day I drag myself to the office chair and force myself to do it instead of walking to the park or eating breakfast for two hours, as I would prefer to. But that struggle – not just the hours spent working, but necessary setting aside of other more immediate pleasures – is part of the output. The hours of my life I spent writing before anyone was willing to read it, let alone pay me for it, all count for something. I feel their presence in the work: they are what make me part of the work and the work part of me, not just some plot or idea.
I don’t doubt that AI will one day be capable of writing a good or even a beautiful novel. But the question is why anyone should want to read such a book, sheer novelty aside, when there are far better books written by real people than one could ever read in one lifetime. The very idea of AI-generated fiction – the belief that it has appeal and value – accords to an understanding of art and writing I fundamentally don’t agree with; it suggests that perfection exists in art, that the nearer to perfection a book comes, the better it is.
I know myself to be deeply imperfect as both a writer and a person. To avoid the shame of those inadequacies overwhelming me into inertia, I have come to see those imperfections, both personal and literary, as intertwined in a way which can be interesting for the work. I have no aspirations of one day making a piece of work that is perfect. But I do have aspirations of making work that is true to myself, that is distinctively Megan Nolan-ish. Not because to write as such is the superior way of writing, but because it is the only metric I have access to and control over. My friend, the non-fiction writer Menachem Kaiser, recently asked an AI bot a prompt to write the same story, first in the style of Menachem Kaiser, then in the style of Megan Nolan. His was wry, witty, erudite; mine emotional. This delighted me, mostly because AI failed to capture either of us, despite superficially understanding our different registers.
This is a relief in logistic terms – I’ve always shuddered at the thought of being a professional whose success is measurable in terms of technical excellence. But aside from letting me off the hook when I am flawed, this perspective accords with what I think is the true power of great writing. Great writing is identified not by its perfect sentences, but from that uncanny feeling of another consciousness rubbing against your own, which can be more intimate and revealing than real-world interactions with others. The writers I love speak to me from the page in whatever way they can, and they tell me, in that particular moment, what it’s like to be alive.