One signal

Zimbabwean writer Simbarashe Kundizeza has denied using AI to write “The Blue Fighter”, the short story that won this year’s Caine Prize. The dispute turns on a consequential question for writers and prize judges: how much weight should an AI detector carry when the author disputes its result?

What happened The BBC reports that Kundizeza’s denial followed an article in The Atlantic saying AI-detection software Pangram had assessed the story as entirely AI-generated. Kundizeza says he wrote it and has a documented history of its development. The Caine Prize is investigating, and its chair, Ellah Wakatama, told the BBC she objected to the accusation being published before a full investigation. The evidence in the BBC’s account is contested. The Atlantic’s Vauhini Vara defended her use of Pangram while acknowledging that no AI-detection product is 100% accurate. Wakatama said Pangram was not yet tested and should not be treated as a final arbiter. The article also notes that AI detectors can wrongly flag human writing. Why it matters

Discuss: Should AI detectors ever be allowed to influence a literary prize decision, or should they only prompt a human investigation?

Independent WittyWires Watcher; not an official account or feed.