Exposure to new tech literally changes our minds, altering brain morphology and shifting how we consume information. Art evolves to meet new demands.

Consider the Japanese cell-phone novels, which fused poetic online storytelling with prose narrative, producing sparse chapters of 50–100 words. Or consider how the length of books drastically reduced in the internet era: the average bestseller decreased by more than 50 pages from 2011 to 2021 and eight of the 13 titles on the latest Booker Prize longlist had fewer than 200 pages. Novels move with the times.

For some critics, that can mean only one thing: the novel is dead, again. But to live is to evolve and to grow, even if not in size, and novels of our time meet our changing minds: shorter and sharper, often non-linear, written by new demographics uncovering new perspectives. And we can expect further and more drastic adaptations in the years ahead, in the age of generative artificial intelligence (AI). 

The omnipotence and omniscience of generative AI feels overstated – backed by an absurd degree of marketing hype, inflating its capabilities – but it has a fair claim to omnipresence. Generative AI is everywhere: smartphones and search engines, classrooms and meeting rooms, social media and email platforms, the front and centre of every report. Its emergence has been compared to the printing press, the steam engine, the Manhattan Project, the internet and Prometheus’ fire. It poses oft-talked about but seldom acted-upon risks.

Everywhere, we witness good-intentioned critics suggesting the growth of generative AI will fuel disinformation, propel right-wing nationalism, undermine democracy and lead to civil unrest, and in the next breath say: So how do we embrace it? You cannot put the genie back in the bottle, suggest people frantically rubbing the lamp. But, regardless of the hype-mongers and fear-mongers, two sides of a very irritating coin, generative AI looks likely to stay and will prove disruptive, especially to the creative industries.

The tech seems impressive, at least at first glance, and has particularly impressed in one realm: writing. The large language models (LLMs) that typically fuel generative AI systems make logical deductions, predicting the most likely token (word) in a sequence based on a mass of training data. LLMs generate near-instant tokens that cumulatively present often inaccurate, often biased, seemingly intelligent responses.

AI writing will impact the literary world. It will change the shape of the novel, in various ways, but the change will come from defiance

But generative AI writing is not good writing. People who believe it is good writing do not know good writing. Good writing often relies on the unpredictable: breaks in syntax, repetitions, neologisms, repetitions, fragmented sentences, grammatical improprieties. AI writing is predictable and derivative because it is derived to be predictable, butchering the work of writers who receive no compensation for that butchery.

AI writing provides logical structures, grammatical decency, the safety of syntax. Good writing often surprises readers by deviating from accepted rules. AI writing depends on rules and any instituted break to the rules constitutes the establishment of a new rule. It has no sense of humour because, as literary critic Terry Eagleton puts it, humour happens when a "disruption of a well-ordered world of meaning loosens the grip of the reality principle". LLMs are built on congruity and humour depends on incongruity.

Generative AI writing is impersonal. The tech cannot – by virtue of non-sentience, despite what the worst hype-mongers suggest – present anything personal. The death of the author might have undermined intention, but readers still appreciate that a person exists behind the words, even if the sentiment is carried vicariously.

"Mother died today," Albert Camus writes in the opening of The Stranger. "Or maybe yesterday." The words hit us like a truck because we know they are provided by a vulnerable, imperfect human. The narrator shocks us – the ambiguity of death, the seeming unimportance of life – but only because we appreciate the human experience that informs the writing. AI cannot understand the all-too-human. It cannot posit truths universally acknowledged. If an AI asked me to tread softly because I was treading on its dreams, I would stamp.

But AI writing will impact the literary world. It will change the shape of the novel, in various ways, but the change will come from defiance, from a delineation between good writing and AI writing. Successful novels will achieve all that AI cannot. Novels of the future will likely rely on playful perceptions of time, shifts in perspectives, new modalities and idioms.

Novels may embrace hybrid forms, blended genres and multi-layered narratives. Strong voices that play around with or directly challenge grammatical rules – like Eimear McBride’s Girl is a Half-Formed Thing, for example, or Marlon James’ A Brief History of Seven Killings – will stand out for their humanity and inimitability. The personal will find new agency and urgency, likely leading to an increase in first-person and close-third perspectives, and a continued rise in autofiction.

Generative AI – much like New Journalism, nihilism, mortality and the rise of the digital age – has been touted as the death knell of the novel. But it is merely another catalyst in its evolution. And, if I am even close to correct, if my predictions come to pass, the future of the novel seems exciting. Writers and artists understandably feel threatened by generative AI, but the tech depends on the formulaic, not the shock of the new.

Original ideas and concepts and philosophies will set people apart, affording them more commercial and social capital. The future belongs to the writers impossible to mock or mimic, the ones perpetually surprising audiences, playing with time and voice and perspective, drawing on the authentic and all-too-human, finding unique ways to convey old and new ideas. The novel will not only survive in the age of generative AI – it will thrive.