Since 2012, a British literary magazine Grant has announced the regional winners of the annual Commonwealth Short Story Prize. This year, however, something went wrong with one of the selections for the prestigious award: It appears to have been written by AI.
Jamir Nazir’s “Snake in the Grove” has many hallmarks of LLM-generated prose—mixed metaphors, anaphora, lists of triplets. (Also, I’m aware this is a list of triplets, and I promise I wrote this post myself, unassisted, because I write all things.) I’ll admit that I wasn’t convinced at first by the claim that Nazir’s story was created by an AI. I know people use LLMs to help them write—or write for them, period—but I was wary of the kind of AI paranoia that had developed among my peers. The em dashes are said to be AI utterances, as are the word “down” and the lists above. Short, punchy sentences too, especially when used to break up a series of longer sentences.
But I, a human being, have certainly used all of the above in my writing before. After all, LLMs are trained in human writing. They reflect what they have been fed. And yet AI-generated prose is terrifying. Something is there, even if you can’t tell what it is right away. If there are specific AI messages and I’m using them right now, how do you know I actually wrote it?
Nabeel S. Qureshi, a former visiting AI scholar at George Mason University’s Mercatus Center, was among the first to point out the suspicious use of AI in Nazir’s story. For Qureshi, the first two sentences were proof enough.
It is said that the grove still hums at noon. Not the neat industry of bees, or the clean scrabble of a saber on vibration, but the sound of a belly—as if the earth had swallowed the scream and held it there.
“In general, AI writing has a particular rhythm that I’ve learned to pick up that is hard to describe,” Qureshi told me via email. “There’s a spectrum from ‘AI helped me edit’ to ‘AI wrote it’ – this case seems like the other end to me, although of course I don’t know for sure.”
The problem is that while the use of AI is widely suspected, none of us know for sure. Commonwealth Foundation CEO Razmi Farook said in a statement that the organization is aware of the allegations regarding artificial intelligence in award-winning stories, including Nazir’s. Farook said that all writers who submitted work for the prize are asked whether they are submitting original, unpublished work, and that all shortlisted writers have personally stated that no artificial intelligence helped them design their stories.
“Until there is a sufficient tool or process to reliably detect the use of artificial intelligence that can also deal with the challenges of working with unpublished fiction, the Foundation and the Commonwealth Short Story Prize must operate on the principle of trust,” said Farook.
Grantfor its part, ran Nazir’s story through Claude “and asked if it was created by AI,” publisher Sigrid Rausing said in a statement. “The response was lengthy and concluded that it ‘almost certainly did not arise without the help of man.'” But Claude is not an AI detection tool, it is a chatbot powered by a large language model. While AI tools are often better than human readers at spotting prose produced by LLMs – or at least those judged by literary awards – Grantthe statement indicated that they went to the source to ask if the story in question was actually created by AI, indicating that perhaps even the magazine itself doesn’t understand how AI works.
“It’s possible that the judges have now awarded the prize for AI plagiarism — we don’t know yet, and we may never know,” Rausing said.
Publications are increasingly tricked into running AI-generated stories, some of them “written” by “authors” who don’t actually exist. There have even been suspicions that Nazir himself is fake – although author Kevin Jared Hosein, a previous winner of the Commonwealth Short Story Prize, confirmed that Nazir is a real person and shared messages he had recently exchanged with Nazir about the suspected use of AI in his story. Nazir also released a poetry collection in 2018. Nazir did not respond The Vergeuser request for comment. In March, Hachette pulled the release of Mia Ballard’s horror novel A shy girl after its author was accused of using artificial intelligence, although Ballard denied using it and blamed a hired editor.
There is also the question of whether there is any acceptable way for authors and journalists to use AI. LLM generated prose is off limits of course, but what about using AI to generate ideas or for research? What about AI transcription services? At what point does relying on these tools mean the work is no longer your own? This week, Polish author Olga Tokarczuk admitted that she uses AI to aid her creative process — the other end of the spectrum of uses for AI that Qureshi alluded to, but troubling to readers who admired the Nobel Prize-winning writer.
“I often just throw an idea into the machine with the challenge: ‘Honey, how can we make this beautiful?'” said Tokarczuk, who was awarded literature’s highest honor in 2018:
“While I am aware of its hallucinations and numerous factual errors in the field of quantitative economics or factual data, I must admit that in the fluid field of literary fiction, this technology is a boon with incredible effect. At the same time, I feel an acute human grievance for an era that is disappearing, never to return. My heart aches at the departure from traditional literature. In all this, I am damned sad for Balzac, Ciorano and the inimitable Nabokov, because despite my enthusiasm, I don’t believe any modern chat can speak in their great way.
Tokarczuk’s comments, delivered in Polish at a recent event in Poznań, had the misfortune of going viral at the same time as the Commonwealth Prize controversy. (We had her remarks translated into English by a human.) But she is far more ambivalent about AI than the headlines surrounding the event would suggest. Tokarczuk clarified her use of AI in a three-point statement shared with Lit Hub, in which she explained that she did not use AI to write her upcoming book, but was using it for “faster documentation and fact-checking,” though she independently verified the information herself.
“Sometimes I am inspired by dreams,” she continued, “but before even this sentence is cornered and torn apart by experts, I hasten to announce that these are my own dreams.”
The uproar over Tokarczuk’s initial comments — and the need for her to explain herself — speaks to a larger, not entirely unjustified, paranoia in publishing about the use of AI. LLM-generated prose may be the new normal, but does anyone want it? Thousands of people threatened to boycott Barnes & Noble after CEO James Daunt said he had no problem selling books written by artificial intelligence as long as the books included a disclaimer that they were not written by a person. Daunt later took back his remarks, but not completely. “Banning books is a clear and present danger, so we are very careful with calls to ban any books,” he said Los Angeles Timeswhile also making sure you “don’t sell AI-generated books impersonating real authors.”
But none of this explains the strange quality of AI-generated work, or what distinguishes bad LLM-produced prose from bad human writing. When I ran Nazir’s story through Pangram, an AI and plagiarism detection software, it came back as 100% AI-generated. According to Pangram, the most obvious narrative was Nazir’s use of triads; the word “stubborn”, which occurs six times more often in AI-generated text than in human-generated text; and the phrase “as if” is five times more likely to occur. But here we have another list of three written by me, the human.
Unsatisfied, I released an unpublished excerpt from my forthcoming book, which I am currently editing, via Pangram. Including one paragraph two triads. (It’s not a very good part of the book, which is why I’m editing it.) Pangram said the passage was 100 percent human-written, which is true, but I still wasn’t satisfied. I read another excerpt—a better one, I think—and he said the same thing. When I started the first chapter Rod a novel by editor Kevin Nguyen, My documentsvia Pangram, the result was the same. Pangram itself ran every Commonwealth Prize winner through its software and found that two of the 2026 winners, as well as the 2025 winner, appeared to have been produced by AI. Human-produced work has a kind of ineffable quality, as does its inverse. Maybe AI-generated prose is like obscenity: You know it when you see it, even if you don’t know why.
About The Author
