Granta published Jamir Nazir's 'The Serpent in the Grove' in May 2026 as a regional winner of the Commonwealth Short Story Prize, and Pangram detection software flagged it as 100% AI-generated. The story opens with mixed metaphors and anaphora — 'They say the grove still hums at noon. Not the bees' neat industry or the clean rasp of cutlass on vibe, but a belly sound — as if the earth swallows a shout and holds it there' — and uses the word 'stubborn' and the phrase 'as if it had,' both statistically overrepresented in LLM output by 6x and 5x respectively.
Granta publisher Sigrid Rausing ran the story through Claude and asked whether it was AI-generated. Claude responded that it was 'almost certainly not produced unaided by a human,' a conclusion that itself demonstrates the gap in detection capability — Claude is a chatbot, not a forensic tool, and asking it to verify its own output is methodologically circular.
Commonwealth Foundation director-general Razmi Farook said all shortlisted writers personally stated no AI was used. The foundation has no technical verification process and relies on self-attestation.
Key facts
- 01Jamir Nazir's Commonwealth Prize story 'The Serpent in the Grove' tested 100% AI-generated by Pangram detection software.
- 02Granta asked Claude whether the story was AI-written; Claude concluded it was 'almost certainly not produced unaided by a human.'
- 03AI-generated text uses the word 'stubborn' 6x more often than human writing and the phrase 'as if it had' 5x more often.
- 04Pangram flagged two of the 2026 Commonwealth Prize awardees and the 2025 winner as AI-generated.
- 05Hachette pulled Mia Ballard's horror novel in March 2026 after AI-use accusations; Ballard denied it.
Pangram ran every Commonwealth Prize winner through its software and found two of the 2026 awardees and the 2025 winner appear to have been AI-generated. The software identifies statistical tells — triads, specific word frequencies, syntactic patterns — that cluster in LLM prose but also appear in weak human writing, making false positives inevitable.
Nazir published a poetry collection in 2018 and is confirmed to be a real person by previous Commonwealth Prize winner Kevin Jared Hosein. Nazir did not respond to press inquiries about the AI allegations.
“Until a sufficient tool or process to reliably detect the use of AI emerges that can also grapple with the challenges pertaining to working with unpublished fiction, the Foundation and the Commonwealth Short Story Prize must operate on the principle of trust.”— Razmi Farook, Commonwealth Foundation director-general
The literary world has faced three AI scandals in recent months. Hachette pulled Mia Ballard's horror novel 'Shy Girl' in March 2026 after AI-use accusations; Ballard denied using AI and blamed a for-hire editor. Nobel Prize laureate Olga Tokarczuk admitted at an event in Poznań that she uses AI with prompts like 'Darling, how could we beautifully elaborate this?' to develop ideas, though she clarified she verifies all facts independently and did not use AI to write her forthcoming book.
Tokarczuk's comments went viral the same week as the Commonwealth Prize controversy, forcing her to issue a three-point clarification. She described AI as 'an asset with unbelievable leverage' in fiction while mourning 'the departure of traditional literature written in isolation over months, a work conceived in the mind of a single conscious individual.'
“It may be that the judges have now awarded a prize to an instance of AI plagiarism — we don't yet know, and perhaps we never will know.”— Sigrid Rausing, Granta publisher
Barnes & Noble CEO James Daunt initially said he had no problem selling AI-written books with disclaimers, then walked back the statement after thousands threatened a boycott. Daunt told the Los Angeles Times the chain would not sell 'AI generated books that masquerade to be by real authors' but stopped short of banning them entirely.
The detection gap leaves publishers, prize committees, and readers without recourse. Pangram flagged a chapter from The Verge editor Kevin Nguyen's novel 'Mỹ Documents' as 100% human-written, correctly, but the tool's reliability on borderline cases is unproven. AI-generated prose exhibits 'an eerie quality' and 'a particular rhythm' that human readers claim to recognize, but no one can articulate the distinction in falsifiable terms.
The industry's reliance on self-attestation creates a selection problem: authors who use AI without disclosure face no consequences unless detection software catches them, and detection software produces false positives on human work with similar statistical properties. The result is a contest between LLMs trained on literary corpora and readers trained on the same corpora, with neither side holding definitive ground truth.
Publishing's AI paranoia has led to the banning of common stylistic choices — em dashes, the word 'delve,' short punchy sentences — as tells, even though human writers have used those devices for decades. The backlash risks penalizing stylistic trends that predate LLMs while missing actual AI use that mimics older literary modes.
The literary world now faces a binary choice: accept that some prize-winning fiction may be AI-generated and judge the work on its merits, or build a detection infrastructure that can parse authorship claims in unpublished manuscripts. The Commonwealth Foundation chose the former, explicitly stating it 'must operate on the principle of trust.' That principle works only until trust becomes statistically unsustainable — and Pangram's 2-in-6 hit rate on recent Commonwealth winners suggests the threshold is near.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
