Caine Prize Confirms Zimbabwean Winner After AI Investigation Clears Author

The organisation behind a prominent African writing prize has officially confirmed Zimbabwean author Simbarashe Kundizeza as this year's winner following an extensive internal review. The inquiry concluded there was no evidence that artificial intelligence was used to generate his successful short story, securing his £10,000 award.

Ellah Wakatama, chair of the board of trustees for the Caine Prize, announced the findings in a video statement. She noted that the review was initiated after a US magazine article questioned the authenticity of the winning entry. Wakatama stated that the evidence gathered clearly indicated the piece was entirely human written.

The initial allegations emerged from an article by journalist Vauhini Vara in The Atlantic, which suggested that AI-detection software had flagged the story about a young boy whose family faces financial difficulties. Those claims sparked a significant social media backlash, with various writers discussing the implications of such accusations within the literary community, particularly regarding non-white authors.

The Caine Prize review confirmed Simbarashe Kundizeza's winning short story was entirely human written.

To address the concerns fairly, organisers adopted a human-centred investigation approach. This included detailed interviews with the author, the judging panel, and the platform where the story was originally published. Additionally, an independent cyber security and digital forensic expert conducted a thorough metadata analysis of early drafts.

Following the comprehensive digital and qualitative assessments, the independent team verified the authenticity of the drafts. Kundizeza previously explained that his creative inspiration came from a poetry reading about fish tanks combined with personal childhood memories from the onset of Zimbabwe's economic hardships in 2002.

Why it matters

The formal clearance of the winning author addresses growing industry-wide anxieties surrounding the reliability of automated detection tools in literary competitions. As major international prizes face increased scrutiny over the potential use of generative technologies, this review establishes a precedent for combining expert digital forensics with qualitative checks to protect authors' reputations.