Another day, another literary scandal involving AI. It has been alleged that the judges of the Commonwealth Short Story Prize have been duped by an author using AI in his winning entry. Jamir Nazir’s The Serpent in the Grove, which won for the Caribbean region, was then published in leading literary magazine Granta, along with other winning entries.
Almost immediately, it attracted accusations of being AI generated. Users on X posted screenshots of reports from AI detection tool Pangram, which claim 100% of the text was AI authored. Of course, the reliability of such tools can’t be guaranteed.
Sigrid Rausing, publisher of Granta, which was not involved in selecting the story, said in a statement that she showed the story to Claude.ai and asked if it was A.I.-generated. “The response was long, concluding that it was ‘almost certainly not produced unaided by a human.’” Again, this may tell us nothing. The New York Times and the Guardian have contacted Nazir for a response to the allegations, but they report he has not responded.
Where does this leave literary prizes, literary writing, and the literary short story?
Razmi Farook, the Commonwealth Foundation’s director general, has defended the integrity of the prize’s judging process. Using AI checkers on unpublished submissions “would raise significant concerns surrounding consent and artistic ownership”. Competitions like the Commonwealth Prize, she notes, must operate on a “principle of trust”.
I agree stories should be received on a principle of trust – but that may be hard to maintain. This year has already seen revelations of significant AI use in a New York Times book review and a debut horror novel. Maybe standards for short fiction will need to shift.
‘Like sunrise over a sink’
The Commonwealth Writers Prize is open to submissions from adult citizens of Commonwealth member states. It typically attracts thousands of submissions each year. Judging panels for each of the award’s regions (Africa, Asia, the Caribbean, Europe and Canada, and the Pacific) decide on regional winners, who are all published in Granta. An overall winner is also chosen.
The Serpent in the Grove was praised by the regional judging panel for its “vivid, lush imagery” and “quiet authority”. It tells the story of a young, unhappily married couple, living in poverty in their rural village, alongside their older, gossipy neighbour. A near – though not entirely accidental – brush with death unexpectedly unites the three of them, creating an awareness of the hidden, subterranean stories that shape their lives and locality.
Writers, literary critics and commentators on X and other platforms have derided what they see as its “obvious” AI writing syntax.
The frequent use of “not x, but y” sentence structures and unusual word repetition have been presented as textual evidence.
The story also includes, to put it charitably, a few choice sentences you’d have to hope were written by a robot.
As Booker Prize winning author Marlon James put it on Facebook:
Forget AI for a minute. A story won an International Competition with a line like this: “The girl smiled like sunrise over a sink.”
But AI usage is difficult to prove. Detection tools are not reliable and have been known to hallucinate their results. Just because the story feels AI generated to many readers, with its somewhat overworked, metaphor-laden prose, hazily dramatised action and slippery refusal to commit to a concrete subject or theme doesn’t necessarily make it the case. It eventually concludes with a string of poetic aphorisms:
A story is a well
It eats sound until somebody throws a rope
If grace is near and hands hold, something breathing comes up.
Ultimately, Rausing has taken a noncommittal stance, stating “perhaps we will never know” if the story was authored by AI.
A new AI milestone?
If the Serpent in the Grove is AI generated, perhaps a new milestone has been reached, with AI now managing to produce the kind of “quotidian, plotless, moment-of-truth revelatory story” that United States author Michael Chabon warned was coming to dominate publications and prize lists.
Regardless of its provenance, I suspect that if you entered the prompt “write a short story that will get at least a regional shortlisting for an international prize” into the AI platform of your choice, something like Serpent in the Grove would be the likely result.
Everything about it, from its frequently solid, if routine, descriptive passages (“Inside, air clung thick as porridge skin: damp earth, woodsmoke, and the sour tang of fermenting cocoa”) to its occasionally sharp observations (“Men who set traps plan for silence, not for the squeal.”) to its outright howlers (“She had the kind of walking that made benches become men”) feels true to the form. And like something that could be created by an authentically human literary mediocrity.
It’s every bit of worthy filler you’ve found padding out the middle of a debut collection published straight out of a prestigious MFA (Master of Fine Arts) writing workshop. It’s every short story you’ve ever skipped past in the New Yorker to get to the film reviews and the caption competition.
What’s the solution?
Certainly, there is still at least a possibility that this small scandal may fall into the time-honoured tradition of the literary hoax.
Despite the bio note by Nazir, a 63-year old resident of Trinidad and Tobago, claiming to be “a prolific poet and author”, readers have had difficulty tracking down any other works. His LinkedIn profile reveals he has previously written on the possibility of AI eliminating jobs and the AI arms race.
The Serpent in the Grove may yet prove to be a parody or a warning.
Advances in AI platforms and composition have often been oddly benchmarked through their engagements with the form of the short story. The launch of ChatGPT saw a major science fiction magazine having to temporarily close submissions due to being bombarded with AI-generated stories.
Last year, Sam Altman of Open AI triumphantly claimed to have been “struck” by how his AI model “got the vibe of metafiction so right” in following his prompt to write “a metafictional literary short story about AI and grief”.
And AI generation can now clearly produce the kind of technically competent creative writing that The Serpent in the Grove exemplifies (the occasional glories of a smile that is “like sunrise over a sink” aside).
Perhaps the solution is to more actively look for new forms and structures for the short story. Or even just make a return to concrete detail, grounded specific action, plot. The measure of a successful story is its capacity to surprise us, however it is created.
This article was originally published on The Conversation. Read the original article.
