Overview
- Online users and writers have applied AI‑detection tools to the published winning stories and accused two regional winners of using machine‑generated text.
- The Commonwealth Foundation said judges avoided detector software for unpublished submissions and that shortlisted authors had twice declared they did not use AI and then reconfirmed those answers when contacted.
- Granta published the winning pieces with an editor’s note saying the speculation is being taken seriously and that the stories will stay online until definitive evidence appears.
- AI detectors and expert readers have offered conflicting readings of the texts, showing that current tools can disagree and are not a reliable sole proof of machine authorship.
- The dispute has highlighted risks to authors’ reputations and prompted calls for clearer prize rules on verification, while organisers warn that uploading unpublished manuscripts to checkers raises consent and ownership concerns.