Caine Prize investigates allegation that winning story was generated by AI
The Caine Prize, the foremost international literary award for short fiction by African authors, has opened an investigation after journalist Vauhini Vara reported in The Atlantic that Zimbabwean writer Simbarashe Steyn Kundizeza used AI to write "The Blue Fighter," his winning story for the 2026 prize, first published in The Shallow Tales Review. Vara, who has long covered AI's effect on writing, fed the story through Pangram, a detection tool she considers more credible than most, which flagged almost the entire text as machine-generated. It is the third major literary prize this year to face such accusations, after the Commonwealth Prize and the Prix Goncourt.
The case is complicated by the unreliability of AI detectors, which studies have shown perform far worse on writing from outside the dominant white Western tradition. A Stanford study found that while detectors were near-perfect on essays by US-born eighth-graders, they classified 61.22% of TOEFL essays by non-native English speakers as AI-generated, and every one of seven tested detectors wrongly flagged 97% of those essays at least once.
Caine Prize board chair Ellah Wakatama said she will approach Kundizeza with compassion and a presumption of honesty, insisting on proof "that does not come from a machine" and promising a human-centred process. The dispute echoes France, where Haitian-born Canadian novelist Thélyson Orélien, accused of using AI, was dropped from the Prix Goncourt longlist and abandoned his book tour — a controversy quickly exploited by the far right, though it also boosted his sales. Author James Folta, writing in Literary Hub, argues the episode illustrates how AI has eroded trust in text in ways that fall hardest on writers from marginalized backgrounds, and that tech boosters show little interest in compromise.
Continue in the app — vote & join in ➔