Writing is a very vulnerable tool. The prize brings those people to the front.
When a short story called 'The Serpent in the Grove' won a place in the Commonwealth Short Story Prize and was published by Granta, it became an unexpected mirror held up to a literary world still learning to see itself in the age of artificial intelligence. Readers flagged suspicious patterns; detection tools cried certainty; experts counseled doubt — and the truth, as is so often the case at the frontier of new technologies, remained stubbornly out of reach. The episode is less a scandal than a symptom: institutions built to celebrate human creativity are now navigating a landscape where the boundaries of authorship have grown genuinely, perhaps permanently, uncertain.
- Readers and researchers flagged 'The Serpent in the Grove' within days of publication, pointing to stilted metaphors and repetitive constructions as signs of machine authorship — and one AI detection tool declared it 100% artificial.
- The accusation arrived as the literary world was already reeling from separate AI controversies involving a halted horror novel and a nonfiction book riddled with fabricated quotes.
- The story's author, Jamir Nazir, whose claimed publication record appears far thinner than his biography suggests, has not responded to any requests for comment.
- Granta and the Commonwealth Foundation offered conflicting defenses — one leaning on Claude's own assessment that the story was likely human-assisted, the other urging caution against the 'hysteria' of false accusations that could devastate vulnerable, emerging writers.
- Experts in linguistics and machine learning warn that AI detection tools are deeply unreliable, especially with creative writing, leaving the question of the story's true origin genuinely unresolved as the prize's final winner is set to be announced on June 30.
A short story called 'The Serpent in the Grove' won a regional prize in the Commonwealth Short Story Prize and was published by Granta, the prestigious British literary magazine. Within days, readers began raising alarms. They pointed to an overabundance of metaphors, strange figurative constructions, and repetitive syntactic patterns as evidence of AI authorship. When researchers ran the text through detection software, it came back flagged as entirely machine-generated.
The accusation landed in an already unsettled literary moment. A horror novel had recently been pulled by Hachette following AI allegations, and a nonfiction author had admitted his book contained AI-generated quotes he had failed to catch. The author of 'The Serpent,' Jamir Nazir, describes himself as a prolific published writer, but his visible body of work is sparse. He has not commented publicly.
Granta and the Commonwealth Foundation responded differently. The foundation's chief executive said the organization trusted its selection process while acknowledging the technological landscape was shifting. Granta's editor took a more guarded stance, noting that when the story was shown to Claude, the AI concluded it was almost certainly not produced without human involvement — while also admitting the prize may have gone to an act of AI plagiarism, and that certainty might never arrive. Granta kept the story online pending definitive evidence, while clarifying its editors had played no role in the selection itself.
Experts cautioned against placing faith in detection tools, which are prone to error — particularly with creative writing that employs unconventional language. A corpus linguistics professor found the story neither obviously human nor obviously artificial. The foundation's chief executive raised a concern that often goes unspoken: false accusations carry real human costs, especially for unpublished writers whose vulnerability is part of what the prize is meant to honor.
The broader literary world's relationship with AI remains unresolved. Some authors experiment openly; others deny any involvement; still others, like Nobel laureate Olga Tokarczuk, find themselves clarifying the precise and limited ways they use it. With roughly 7,800 submissions received this year and a final winner to be announced on June 30, the question of whether 'The Serpent in the Grove' will be crowned — or disqualified — hangs unanswered over an institution still searching for the language to describe what authenticity means now.
A short story called "The Serpent in the Grove" won a place in the Commonwealth Short Story Prize and was published online by Granta, a prestigious British literary magazine known for publishing Kazuo Ishiguro, Zadie Smith, and Salman Rushdie. Within days, readers began raising questions. They spotted what they saw as telltale signs of artificial intelligence: excessive metaphors and similes, nonsensical figurative language (one line in particular—"it had the kind of gait that made benches turn into men"—drew particular scrutiny), and a reliance on the "not X, but Y" construction. When some readers, including Ethan Mollick, a researcher at the Wharton School, ran the story through AI detection tools like Pangram, the software flagged it as 100 percent machine-generated.
The accusation landed at a moment when the literary world was already on edge. In March, Hachette Book Group had halted publication of a horror novel called "Shy Girl" after widespread online claims that its author, Mia Ballard, had relied heavily on AI. Ballard told the Times that someone she hired to edit a self-published version had used AI, but that she herself had not. Days before the "Serpent" controversy erupted, the Times also reported that author Steven Rosenbaum had included several fabricated or misattributed quotes generated by AI in his recent nonfiction book about artificial intelligence, "The Future of Truth." Rosenbaum acknowledged the book contained "a handful of improperly attributed or synthetic quotes" and said he had launched his own investigation.
The author of "The Serpent in the Grove," Jamir Nazir, describes himself in his Granta biography as a prolific poet and author with published books and forthcoming works. His actual publication record appears thin—among his few visible works is a self-published poetry collection titled "Night Moon Love: Poems for All Who Have Loved or Dreamed of Love." Nazir did not respond to requests for comment.
Granta and the Commonwealth Foundation, which administers the prize, took different public stances. Razmi Farook, the foundation's chief executive, told the New York Times in a video call that the organization had taken the concerns seriously and examined its selection process. "We trust in the rigor of our process, but we are aware that this is an evolving technological environment," he said. Granta's editor, Sigrid Rausing, adopted a more defensive posture. She said the magazine had shown the story to Claude.ai and asked whether it was AI-generated. The tool responded with a lengthy conclusion that the story was "almost certainly not produced without human assistance." Rausing added: "It may be that the judges have now awarded a prize to a case of AI plagiarism; we still don't know, and perhaps we never will." On its website, Granta published a statement distancing its editors from the prize's selection process, noting they had not been involved in choosing the stories, only in editing them after receipt. The magazine said it took seriously the suggestion that writers had submitted inauthentic material, but would keep the stories online pending definitive evidence.
The controversy exposed a fundamental problem: it is genuinely difficult to determine whether a piece of writing was created by a human or a machine. The AI detection tools that flagged "The Serpent in the Grove" as entirely artificial are themselves unreliable. Nicholas Andrews, a principal researcher in computer science at Johns Hopkins University whose work focuses on AI and machine learning, cautioned that these detectors "make many errors, especially in creative writing that uses unusual constructions that may not be typical of what Pangram has trained its system on." Jack Grieve, a corpus linguistics professor at the University of Birmingham, noted that without controlling for variations in dialect, subject matter, genre, and prompts, relying on AI detection tools is risky. In his assessment, the short story was neither obviously AI-generated nor obviously human-written.
Farook acknowledged the limits of current technology. "The technology is not yet up to the task, and because of its imperfections, it leads people to think things are black and white when they may not necessarily be," he said. He cautioned against "an impulsive reaction to the general hysteria around these topics right now." He also raised a concern that has received less attention: the risk of false accusations to writers themselves. "These are people with personal stories, writers facing their own personal challenges, and writing is a very vulnerable tool," Farook said. "The prize brings those people to the front. They have to face something new. Often they are unpublished writers, so there is a duty of care and consideration we must keep in mind."
Meanwhile, the literary world's relationship with AI remains fractured and evolving. Some authors have been openly experimental. Coral Hart self-published more than 200 romance novels last year with help from Anthropic's Claude. Nobel Prize-winning Polish author Olga Tokarczuk caused a stir this week when she told an audience in Poland that she had used AI while writing her latest novel—though she later clarified through her publisher that she uses AI only as a research tool for faster documentation and verification, not for actual writing. The Commonwealth Short Story Prize received roughly 7,800 submissions this year. Of the five stories selected as regional winners, one will be named the overall winner on June 30. The question of whether that winner will be "The Serpent in the Grove" remains unresolved.
Notable Quotes
We trust in the rigor of our process, but we are aware that this is an evolving technological environment.— Razmi Farook, Commonwealth Foundation chief executive
These are people with personal stories, writers facing their own personal challenges, and writing is a very vulnerable tool. There is a duty of care and consideration we must keep in mind.— Razmi Farook, on the risks of false accusations
The Hearth Conversation Another angle on the story
So the story was flagged by readers, not by the judges themselves. What made people suspicious in the first place?
The writing had patterns that readers associated with AI—too many metaphors stacked together, some phrases that didn't quite make sense, a repetitive structure. One line about benches turning into men got particular attention. But here's the thing: those patterns show up in human writing too, especially experimental or stylized work.
And when they ran it through detection software, it came back as 100 percent AI. Case closed?
That's what you'd think, but the experts say these tools are unreliable, especially with creative writing. They're trained on certain kinds of text, and if you write in an unusual way, the detector gets confused. It doesn't mean the story is definitely AI or definitely human.
So we might never know for certain.
Probably not. And that's what makes this hard for everyone—the author, the judges, the magazine. You can't prove a negative.
What about the author himself? Has he said anything?
Nothing. He hasn't responded to requests for comment. His publication history is thin—mostly a self-published poetry book. That made people more suspicious, but it doesn't prove anything either.
The prize organizers seem caught between defending their process and admitting they can't really verify anything.
Exactly. They're saying their selection was rigorous, but also acknowledging that the technology for detecting AI is still evolving. What worried me more was the point about false accusations—if you accuse a writer of plagiarism and you're wrong, you've damaged a person's reputation and career. That matters.