Commonwealth Short Story Prize winners face AI allegations
The controversy centres on Jamir Nazir’s Caribbean-winning story, “The Serpent in the Grove,” which AI detection tool Pangram flagged as 100 per cent AI-generated.
Three of the five regional winners of the 2026 Commonwealth Short Story Prize are facing allegations of using generative artificial intelligence to write their entries. The controversy centres on Jamir Nazir’s Caribbean-winning story, “The Serpent in the Grove,” which AI detection tool Pangram flagged as 100 per cent AI-generated. Two other winners, John Edward DeMicoli and Sharon Aruparayil, also faced similar allegations. The Commonwealth Foundation stated it does not use AI checkers due to consent and ownership concerns, relying instead on the principle of trust. Granta, which hosts the winning stories, noted its own review was inconclusive and added a disclaimer to the prize page.
The allegations have sparked intense scrutiny from the literary community, with many readers and writers expressing dismay that the prize jury may have overlooked signs of inauthentic authorship. Jamir Nazir’s story, “The Serpent in the Grove,” was the primary focus of the backlash after researcher Nabeel S. Qureshi highlighted specific syntactic markers on social media platform X. WIRED independently confirmed that Pangram flagged the text as entirely AI-generated, noting that third-party analysis has consistently determined the tool to be the most accurate with a near-zero rate of false positives.
Nazir, a writer from Trinidad and Tobago, did not return requests for comment via his Facebook page. While some speculation suggested Nazir could be an entirely AI-created persona, a 2018 article in the Trinidad and Tobago edition of the Guardian suggests he is a real person, having previously self-published a poetry collection titled Night Moon Love. Despite this, posts on his Facebook page and a LinkedIn profile associated with him also scan as AI-generated on Pangram.
The Commonwealth Foundation, a London-based nongovernmental organisation, defended its judging process as robust but explained its reliance on trust. Razmi Farook, the Foundation’s Director-General, stated that the organisation does not use AI checkers because supplying unpublished original work to such tools raises significant concerns regarding consent and artistic ownership. Farook emphasised that all shortlisted writers have personally confirmed that no AI was used in their submissions, and the entry rules require entrants to certify their work as their own original creation.
Granta, the UK literary magazine that has hosted the winning submissions since 2012, conducted its own review using Anthropic’s Claude agent but found the results inconclusive. Publisher Sigrid Rausing acknowledged the allegations but noted that Granta editors have no control over the selection of the prize stories. A disclaimer has been added to the prize page on Granta’s website, stating that the stories will remain until the Commonwealth Foundation reaches a definite conclusion.
Beyond Nazir, Pangram flagged John Edward DeMicoli’s winning story for the Canada and Europe region as fully AI-generated, and Sharon Aruparayil’s entry for the Asia region as partly AI-generated. Neither author immediately returned requests for comment via social media. In a further development, Jamaican author and judge Sharma Taylor was accused of using AI to craft the descriptive blurb for Nazir’s story, with Pangram evaluating her text as AI-assisted.
The incident reflects a broader trend of AI-related controversies across the publishing and academic sectors. Recent weeks have seen author Steven Rosenbaum acknowledge AI-hallucinated quotes in his new book, Nobel laureate Olga Tokarczuk admit to using large language models in her creative process, and arXiv announce one-year bans for authors who fail to catch erroneous AI-based content.
