In May 2026, a story called "The Serpent in the Grove" was published in Granta, one of the most prestigious literary magazines in the English-speaking world. It had been selected as the Caribbean regional winner of the Commonwealth Short Story Prize — a competition that received 7,806 submissions from writers across the former British Commonwealth, the second highest entry count in the prize's history. The judging panel was chaired by Louise Doughty, an award-winning English novelist.
Within days of publication, readers on Bluesky and X began raising questions about whether the story had been written by a human at all.
7,806 Entries. Five Winners. One Controversy.
The accusations centered on what critics described as textual markers associated with AI-generated writing: sentences structured with items grouped in threes, recurring use of "not x, but y" constructions, and what some identified as pangrams — phrases that use every letter of the alphabet, a pattern AI systems sometimes produce. Ethan Mollick, a Wharton professor who studies AI's effects on creative fields, broke down the evidence on Bluesky in a detailed thread, calling it "a Turing Test of sorts."
The story follows a rum-drinking farmer who encounters an enchanted grove in rural Trinidad. The judges had praised its "precise yet richly evocative" language and selected it from successive rounds that reduced 7,806 entries to 200, to 25, to five. A panel of accomplished literary figures had read, discussed, and chosen it — and none of them flagged any concern.
The Markers Critics Found — and the Author's Explanation
Jamir Nazir, the story's author, strongly denied using AI. His explanation for the stylistic patterns critics identified: he has chronic health conditions that make desk-bound typing impossible, so he composes entirely on his Android phone using speech-to-text software. The rhythm of dictated prose, he suggested, may produce different sentence structures than keyboard writing.
The Commonwealth Foundation reviewed the available evidence and stood by Nazir and the other regional winners. In a formal statement, it affirmed confidence in its contributors and its judges. All shortlisted writers had confirmed no AI was used.
Granta's publisher, Sigrid Rausing, acknowledged that the judges "might have rewarded an instance of AI plagiarism — we don't yet know, and perhaps we never will know."
Why Nobody Can Prove It
This is where the story becomes genuinely alarming — not as a specific dispute about one story, but as a preview of a category of problem that literary culture, and culture more broadly, has not yet developed tools to address.
There is currently no reliable method for proving that a piece of text was or was not generated by AI. Detection tools exist, but they produce false positives with meaningful frequency and have been shown to misidentify human-written text as AI-generated, particularly when the human uses dictation, speech-to-text, or unconventional compositional methods. Nazir's explanation is plausible. So is the alternative. Neither can be verified.
As author Tony Tulathimutte observed in a widely shared Instagram post, the outcome "should not disturb anyone who is at all familiar with what awards judges tend to favor" — a burn that lands differently than intended, raising the possibility that literary awards have always been susceptible to certain kinds of mimicry.
Granta's Response and What It Means for Literature
Granta's response was institutional and swift. The magazine announced it would no longer publish Commonwealth Short Story Prize winners, citing its inability to maintain editorial control over external partnerships. The shortlisted stories were kept online "in the public interest." Granta would step back entirely.
The Commonwealth Foundation announced a full review of its judging process to address AI-related concerns in future competitions. The prize organizers did not respond to The Guardian's request for comment.
What no one said explicitly, but what the response implies: if you cannot tell whether a winning entry was written by a human, the prize as an institution loses its function. It was designed to identify and celebrate human creative achievement. That function depends on being able to verify what you're celebrating.
An AI scholar called the controversy a "major milestone" — not because AI had definitively won, but because the ambiguity itself was now the story. We have arrived at the moment where the question "did a human write this?" cannot be reliably answered, even by experts, even with financial stakes attached, even in one of the most celebrated literary magazines in the world.
