Literary Prize Questioned Over Possible AI Authorship
A short story that won the Commonwealth prize for Caribbean writers may have been generated by artificial intelligence, according to accusations that have prompted the prize foundation and its publisher to acknowledge they cannot determine the work's true authorship.
The Serpent in the Grove, published in Granta magazine on Saturday, was written by Jamir Nazir, reportedly a 61-year-old from Trinidad and Tobago with limited prior publications. Within hours of publication, critics and academics flagged the work as potentially AI-generated.
An AI detection platform called Pangram flagged the story as machine-written. A professor at the University of Pennsylvania called it a "Turing test of sorts" on social media. Other commentators cited syntactical patterns common to AI output, particularly repeated "not x, but y" sentence structures.
Sigrid Rausing, publisher of Granta, said the magazine had tested the story using Claude, an AI tool, which returned ambiguous results. "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," she said.
What the Prize Foundation Says
The Commonwealth Foundation did not use AI detection tools during judging, citing concerns about consent and artistic ownership when handling unpublished work. All entrants signed declarations stating their submissions were their own and that no AI was used.
Razmi Farook, the foundation's director general, said: "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process to reliably detect the use of AI emerges that can also grapple with the challenges pertaining to working with unpublished fiction, the foundation and the Commonwealth short story prize must operate on the principle of trust."
The foundation acknowledged that AI detection tools are "not unfailing and infallible."
The Broader Problem
This case is part of a wider pattern. The New York Times terminated a freelance journalist in March after he admitted using AI to write a book review. Hachette cancelled the release of a debut horror novel over concerns it was written partially with AI.
AI detection has become a cottage industry. Research shows these tools perform well in controlled tests but face a fundamental challenge: as detection improves, writers and AI systems will adapt, creating what researchers call "a continuous technical arms race."
The irony, Rausing noted, is that AI itself may be the most reliable tool available for identifying AI-generated text - yet even that produces uncertain results.
Granta said it will keep the story on its website pending a definitive conclusion from the Commonwealth Foundation.
For writers, the situation highlights the stakes of AI in publishing. As detection tools proliferate and improve, the question of authorship - and how to prove it - remains unsettled. AI for Writers and Generative AI and LLM resources explore how these technologies work and their implications for creative work.
Your membership also unlocks:
