The rain came before the monsoon. It came at half past five in the morning and it came hard, and the earth, which had been waiting a long time, took it. The streets filled quickly.”

If Ernest Hemingway had read that short passage while wasting an afternoon at El Floridita in Old Havana, he might have scratched his head at a style so lean yet forceful. Or perhaps it reads like a lost slice of Khushwant Singh. It has the kind of soulful texture you expect from a seasoned human writer.

Yet, given the speed and scale of artificial intelligence’s advance into the literary world, the selection committee of the Commonwealth Short Story Prize and the folks at the British literary journal Granta could well be forgiven for the fiasco that followed the announcement of this year’s winning entries.

Since 2012, Granta, which has featured work by literary giants like Kazuo Ishiguro and Salman Rushdie, has published the regional winners of the competition. Each year, the London-based non-governmental organisation Commonwealth Foundation awards its short story prize to one writer in each of five regions: Africa, Asia, Canada and Europe, the Caribbean and the Pacific. An overall winner is then selected from that shortlist.

This year, Jamir Nazir’s The Serpent in the Grove found itself in the eye of the storm. After its publication on the Granta website, readers quickly pointed out that the style mirrored LLM-generated prose.

Nabeel S. Qureshi, former visiting scholar of AI at the Mercatus Centre at George Mason University in the US, posted on X — “Well, this is a first: a ChatGPT-generated story won a prestigious literary prize. ‘Not X, not Y, but Z’ sentences everywhere, the ‘hums’ trope, and plenty of other obvious markers of AI writing. ” Others quickly chimed in, sharing screenshots of the AI-detection tool Pangram flagging the piece as machine-written.

The problem runs deeper than a single story.

Currently, no AI tool can definitively prove the human authenticity of a piece of writing. Commonwealth Foundation director-general Razmi Farook noted that the organisation is aware of the allegations, adding that they are “conscious that this is an evolving technological environment”.

Sigrid Rausing, the publisher of Granta, which merely prints the winning entries and has no say in the selection process, released a statement of her own, “We showed Claude.ai the story and asked whether it was AI-generated. The response was long, concluding that it was ‘almost certainly not produced unaided by a human’.”

The fallout from this debate has cast a shadow over the entire competition. All of this year’s winners now have to bear the brunt of a rather sour disclaimer added to their story introductions, noting that there has been speculation that some entries may have been at least partially AI-generated.

Naturally, writers are anxious. A study conducted for the University of Cambridge’s Minderoo Centre for Technology and Democracy surveyed 258 published novelists and 74 industry figures. A striking 51 per cent of novelists believed AI would eventually replace their work entirely.

Many reported their writing had already been used without permission to train large language models, while 39 per cent noted a drop in income due to generative AI.
These anxieties are already playing out in the publishing world with high-profile casualties. Hachette Book Group recently discontinued the popular horror novel Shy Girl by Mia Ballard in the UK and cancelled its US launch following online allegations of heavy AI reliance. Ballard told The New York Times that she wasn’t to blame; rather, an acquaintance she hired to edit a self-published version of the book had used generative AI.

Similarly, author Steven Rosenbaum was found to have included fictitious or misattributed AI-generated quotes in his recent non-fiction book about artificial intelligence The Future of Truth. Rosenbaum
is now working with
editors on a full “citation audit”.

Despite the tension, some authors are embracing the tech cautiously. Polish author Olga Tokarczuk, winner of the Nobel Prize in Literature, clarified her stance during a cultural event, stating she treats AI simply as a tool for faster research and fact-checking, not for the actual writing.

Meanwhile, Barnes & Noble CEO James Daunt sparked controversy by stating he had no problem selling AI-written books, provided they didn’t masquerade as human work. Following boycott threats, Daunt clarified to The Los Angeles Times that while he opposes book banning, the shop remains vigilant against AI books pretending to be by real authors.

Ultimately, authors are in a bind. Even using AI for minor inspiration can backfire, much like filmmaker Martin Scorsese faced criticism just for using AI to quickly generate storyboard images.

There is a real danger that AI will flood the Internet with mediocre text, creating a monoculture. It is a risk that somewhat mirrors the Irish potato famine of 1845, where relying on a single crop proved disastrous for a dependent population. When we asked Claude to suggest names for a sci-fi protagonist, it offered predictably cliche choices: Kael Morrow, Seren Voss, Idris Talyne, Zephyra Oult and Cassian Drek.

Now imagine a million articles filling the Internet, all grown from that exact same creative soil.