TimeBinder

Start →

AI Accusation Casebook

Commonwealth Short Story Prize 2026: the AI allegations and the review

In May 2026, Jamir Nazir’s story “The Serpent in the Grove”, the Caribbean regional winner of the 2026 Commonwealth Short Story Prize, was accused on social media of showing signs of AI writing. AI detectors gave conflicting results. The Commonwealth Foundation reviewed working drafts, time-stamped documents and notes, and on 22 June 2026 said: “we are satisfied that AI was not used to write the winning stories.” The story went on to win the overall prize, and Granta, which had published it, ended its external publishing partnerships. This entry records what was reported and who said it. It doesn’t offer a view on what happened.

Table of Contents

Commonwealth Short Story Prize 2026: case summary

Field Entry
Case 2026-001
First reported (fetched sources) 20 May 2026 AEST, The Guardian [N2]
Field Fiction (short story)
Named party Jamir Nazir, Trinidad and Tobago (prize winner, publicly reported)
Who raised it Social media users and some critics, on X and Bluesky [N4]
Detectors named in reports Pangram, Grammarly, GPTZero, QuillBot, as reported [N3]
Evidence examined Working drafts, time-stamped documents and notes, plus discussions with the winners [N1]. Not public.
Outcome status Finding published
Outcome “we are satisfied that AI was not used to write the winning stories” (Commonwealth Foundation, 22 June 2026) [N1]
Last reviewed 3 November 2026

What was alleged

According to the Guardian, the story “went viral after being named as a regional winner in mid-May, with critics on X and Bluesky claiming it showed “obvious markers” of AI use” [N4]. Critics pointed to style features: the Guardian reported that the story “includes multiple “not x, but y” constructions and lists of three, which some consider to be signs of AI use” [N4].

Some people also ran the story through AI detectors. The Independent reported that two tools identified it as “100 percent” AI-generated, while another “classified it as “entirely human”” and a fourth gave a “zero percent likelihood” result [N3].

The Independent also reported that the allegations later extended to two other regional winners’ stories [N3]. This entry covers the Caribbean story only.

What each side said

The prize body. In May, the Guardian reported that the Commonwealth Foundation said all entrants “had avowed that their submissions were their own work and “personally stated that no AI was used”” [N2]. The Foundation said it didn’t put unpublished work through AI checkers because it “would raise significant concerns surrounding consent and artistic ownership” [N2]. At that stage, the Foundation and Granta “had not reached a conclusion as to whether they were true” [N2].

The writer. Nazir did not comment to the Guardian [N4]. In a film released by the Foundation, as reported by the Guardian, he said he wrote “six or seven drafts” and used speech-to-text software. He explained that he could see only a few lines on his phone screen at a time, so he would perfect each line before moving on, which is how the story ended up “highly polished” [N4].

The publisher. Granta said the speculation was about whether “one or more of the stories may have been at least partially AI-generated, accusations that were strongly rejected by the authors” [N7]. It also said “Granta editors were not involved in the Commonwealth regional prize juries” [N7].

Evidence examined

The Foundation’s 22 June statement says: “We held detailed discussions with all regional winners about their creative process, and they collaborated fully in our review. We also examined evidence related to the development of their stories, including working drafts, time-stamped documents and notes.” [N1]

Speaking to the Guardian later, the Foundation said “we asked our winners to show their working drafts, outlines, the evidence of an artistic journey” [N4].

None of the sources fetched for this entry publish the review material.

Outcome

On 22 June 2026, the Commonwealth Foundation said: “After a thorough consultation with our judges and careful consideration of all available information, we are satisfied that AI was not used to write the winning stories. Therefore, we will proceed with the regional winners selected by the independent judging process.” [N1] That is the Foundation’s finding, quoted as the Foundation’s words.

Granta’s note on the story page says: “For the sake of our own editorial integrity, the Granta Trust board has now taken the decision that we will no longer engage in external publishing partnerships where we have no editorial control.” [N7] The Foundation said: “While we respect their decision, we remain confident in the integrity of our judging process and this review.” [N1]

On 30 June the story was named the overall winner [N4][N6]. The judging chair called it “an original, poetic and deeply moving story” [N4]. The Guardian reported that initial social media reactions to the announcement “were negative” [N4]. The Independent reported prize money of £2,500 for each regional winner and £5,000 for the overall winner [N3]. The Foundation said “Further strengthening our processes is now an immediate priority.” [N1]

What this case does not tell us

  • It doesn’t show what the review material contained. The drafts, documents and notes aren’t public, so readers can’t assess them.
  • It doesn’t settle what detectors can do. Different tools gave different results on the same story. That is one reported example, not a test.
  • It doesn’t tell us how other prizes or publishers will act. The Foundation and Granta took different approaches.
  • It isn’t our finding. The outcome is the Foundation’s, in its words. We don’t add a view on whether AI was used.

What a dated record would and wouldn’t have shown

The Foundation said it examined “time-stamped documents” [N1]. This is where a dated record of a file fits, and where it stops.

With TimeBinder®, a writer can TimeBind a draft as they go. Each file’s SHA-256 fingerprint is recorded on the Bitcoin blockchain, which shows the file existed no later than the recorded time. Files are hashed in the browser and never uploaded. Later, anyone with the file can check it still matches its fingerprint. See how a TimeBind works.

A series of TimeBound drafts would have shown:

  • that each draft file existed no later than its recorded time
  • the order in which those files were recorded

It would not have shown:

  • who wrote the drafts
  • whether AI was used at any stage
  • that the contents are original or true
  • how the Foundation, Granta or anyone else would judge them

There are two ways to TimeBind:

  • Pooled (1 Token): Your fingerprint is grouped with other users’ in one shared Bitcoin entry, usually recorded within 24 to 48 hours, and often sooner, depending on Bitcoin network activity. It’s verified through TimeBinder.
  • ASAP (3 Tokens): It gets its own Bitcoin transaction, sent straight away, and can be verified independently from the certificate and the Bitcoin blockchain.

The ASAP vs Pooled verification guide explains the difference. For a free daily habit, Draft Streak logs your word count and a fingerprint of each day’s draft in your browser. It uses your device clock, so on its own it isn’t a timestamp. More free tools for creators are on one page.

If you’re facing an accusation yourself, our guides for an author accused of using AI and a calm reply to an AI accusation cover next steps. Also see whether file metadata can show AI use and what AI text watermarks mean for writers.

When TimeBinder isn’t the right tool

  • You need someone to assess whether AI was used. A TimeBind can’t do that, and neither can we.
  • The drafts are already gone. A TimeBind records files you have now; it can’t date earlier versions.
  • A prize or publisher has its own process. Follow their process. Service terms for TimeBinder are in the TimeBinder Terms and Conditions.

Frequently Asked Questions

  1. What happened at the 2026 Commonwealth Short Story Prize?

    The Caribbean regional winner, “The Serpent in the Grove” by Jamir Nazir, was accused on social media of showing signs of AI writing. The Commonwealth Foundation reviewed drafts, time-stamped documents and notes, and said on 22 June 2026 that it was satisfied AI was not used to write the winning stories.

  2. What did the Commonwealth Foundation look at?

    According to its statement, it held detailed discussions with all regional winners and examined working drafts, time-stamped documents and notes. In May, the Guardian reported that the Foundation hadn’t put unpublished work through AI checkers, citing concerns about consent and artistic ownership.

  3. Did AI detectors agree about the story?

    No. The Independent reported that two tools identified it as fully AI-generated, while two others rated it as human-written.

  4. Why did Granta end its partnership?

    Granta said it decided, “for the sake of our own editorial integrity”, to stop external publishing partnerships where it has no editorial control. Its note describes the allegations as speculation that the authors strongly rejected.

  5. Would a timestamp have shown whether AI was used?

    No. A timestamp shows a file existed no later than the recorded time. It doesn’t show who wrote the file, whether AI was used, or how a prize or publisher would judge it.

Sources

All sources last checked 3 November 2026 (AEST). Named in this entry and want to add a statement or a correction? Corrections: support@timebinder.io

Keep your own process record

If you want a dated record of your drafts as you go, start with a Pooled TimeBind. See plans and pricing