Allegations that an award-winning short story was written by AI have sparked a row over the limits of prize judging and how the publishing industry can respond.

On May 22, IT outlet The Verge reported that Zamir Nazir's short story "The Serpent in the Grove," published by British literary magazine Granta, is under suspicion of resembling text written by a large language model.

One of the first to raise the allegation was Nabil S. Qureshi (나빌 S. 쿠레시), a former visiting AI scholar at George Mason University’s Mercatus Center. He said the first 2 sentences alone raised strong suspicion. Qureshi said AI writing has a distinctive rhythm that is hard to explain, and the work reads closer to something written by AI than something merely edited with AI’s help. He added that he could not be certain.

The problem, he said, is that there is no clear way to prove such suspicions. Razmi Farook (라즈미 파룩), secretary-general of the Commonwealth Foundation, said the foundation checks with entrants that their submissions are original and unpublished, and finalists also said directly they did not use AI to draft their work. He said that until reliable detection tools or procedures emerge, the foundation and the prize have no choice but to operate on a principle of trust.

Granta’s response also added to the controversy. Publisher Sigrid Rausing (지그리드 라우징) said she put the work into Anthropic’s Claude and asked whether it was AI-generated, and received an answer suggesting it was almost certainly not produced without human help. Claude, however, is not an AI detection tool but a chatbot based on a large language model. Rausing said judges may have awarded a prize to an AI plagiarism case, but it is not possible to know now and it may never be known.

Similar confusion is continuing in publishing. Cases are increasing in which AI-generated writing is published under author names that do not exist, and in March the publication of Mia Ballard’s horror novel "Shy Girl" was cancelled over allegations of AI use. Ballard denied it and argued an outside editor was responsible. Nazir was also suspected of not being a real person, but previous winner Kevin Jared Hosein (케빈 재러드 호세인) confirmed that he is real.

It is also unclear how much AI writers are allowed to use. Olga Tokarczuk (올가 토카르추크) said at a recent event that she uses AI in her creative process. She later said in a separate statement that she did not write her next work with AI, and uses it only to organise materials more quickly and to check facts, while verifying information herself.

Detection tools have also shown their limits. Pangram classified Nazir’s work as 100% AI-generated, and judged that 2 winning works from the 2026 Commonwealth short story prize and 1 winner from 2025 may also have been written by AI. It classified other manuscripts written by people as 100% human-written. The literary world is continuing the debate without clearly distinguishing what is written by AI and what is written by humans.