Wharton professor Ethan Mollick said he “roughly” ran the story through an AI detection tool. The tool returned what it described in a post on Bluesky as “100% red flags,” adding, “If you know, come on.” That single post is what turned the regional award win into a months-long, still-unsettled controversy over whether Jamir Nazir actually wrote the story that won him the Commonwealth Short Stories Award.
What actually happened
Minister was named the Caribbean regional winner of the Commonwealth Short Story Prize 2026 for his story The Serpent in the Shelter, one of five regional winners named that year alongside Africa, Asia, Canada and Europe and the Pacific. The prize is judged by a panel chaired by writer Louise Doughty and publishes the winning entries. Grant. Within weeks of the announcement, Mollik’s claim to detect artificial intelligence using a tool called Pangram, which its developers say is about 99% accurate, has placed Minister’s win at the center of a controversy he did not expect the award to control.
What does evidence actually consist of?
The accusation rests on two separate levels. The first is Pangram’s automated flag, which, according to the report Weekwas not limited to Nazir’s story; the same tool also recorded two other winning entries from the same year’s award. The second point is more telling: Readers who began to review the story after Mollick’s writing pointed out certain stylistic tics, including vague reinforcements and a workshop-polished tone that some have described as reading like a parody of MFA, which are conventional signs of AI involvement. Neither strand is a confirmed finding. The output of an AI detection tool is a flag of probability, not determination, and a stylistic similarity to workshop prose is a common feature of modern literary fiction not invented by AI.
What does Nazir say?
Nazir denied that he wrote the story with AI, saying it drew directly on his own childhood memories. GrantThe publisher of award-winning stories said the claim was being investigated, and confirmed one detail. Free Press‘s provides information on the jury structure of the award.
What the Commonwealth Fund actually said
The idea that the awards committee says nothing is outdated as of this writing. The Commonwealth Fund own statements The show’s CEO, Razmi Farook, issued a public response on May 19 defending the judging process, saying the Foundation did not conduct AI detection checks on unpublished artistic submissions, citing consent and ownership concerns and noting that each shortlisted writer had personally denied using AI. Faruk also called the means of detection “infallible” and said that the Fund operates on the “principle of trust”.
Three days later, on May 22, the Foundation issued a second statement saying it “takes the allegations seriously” and is reviewing “all available evidence,” but that it does not intend to reverse or change Nazir’s win. A further update on the progress of the review was provided in June. What hasn’t happened, months later, is a decision: no refund, no confirmation, no public timeline for when the investigation will conclude.
Why silence matters is narrower than it seems
The fund did not remain silent, but it remained uncertain, a different and in some ways more open position. Defending due process while simultaneously reviewing “all available evidence” is not so much a contradiction as an organization trying to hold two things at once: trusting a system built on trusting its writers, and acknowledging that trust itself is now something under public scrutiny. The award, built on the assumption that the shortlisted writer’s word is enough, has become a direct undertaking that a growing number of outside observers no longer consider sufficient, and the Foundation’s own statements show that it is undecided on which side of the gap it will stand.
The same trust problem is manifesting itself in publishing this year
The Commonwealth Fund’s position that it relies on writers’ words over automated detection because detection tools aren’t perfect echoes a rationale that has emerged elsewhere in the publication this year: trade bodies offer “human author” labels that check memberships, not manuscripts, and retailers offer disclosure standards that check a book’s provenance.
Each institution has independently come to the same practical position that verifying authorship at the actual text level is not something they can currently do at scale, and each has substituted some form of self-reporting or institutional trust in its place.
This replacement version of the Commonwealth Fund is now the most publicized option because, unlike a retail shelf or a trade body’s membership list, the entire value of a literary prize depends on the validity of the claim currently in dispute.
What remains unresolved
Months after the indictment, Nazir’s name is still attached to the win, the investigation is still open, and no public evidence has emerged beyond the possible flag of the detection tool and the stylistic impressions of readers. This is a failure of process, rather than an accurate reflection of how little any body, including award committees, should do about an authorship charge once the decision has already been made public.
Whether the review changes anything will say more about how the literary establishment plans to handle the next version of this controversy than it does about settling the current controversy. A finding against Minister would be the first time a major prize has reversed a victory over an AI authorship dispute; a finding that clarifies it would do little to resolve the broader question because the means used to bring the charge remain unproven, as the Foundation itself says.
For now, the only settled fact is one that Foundation himself keeps repeating: he doesn’t know yet, and neither do outsiders reading the story based on the same probabilistic tools and stylistic guesswork that fueled the controversy in the first place.
It is worth keeping this distinction straight for months in an argument that produces nothing stronger than suspicion on the one hand and denial on the other.
