AI Is Already Winning Literary Prizes, and the Only Way We Can Prove We’re Human Is to Write Badly

A Commonwealth Prize Caribbean-winning story published by Granta drew online suspicion that it may have been written by AI.

Granta asked Claude to assess it, and the model denied the text was AI-generated.

The case highlights how AI-writing style can seem obvious to readers, while automated detectors often miss the mark.

Tools like GPTZero, Originality.ai, and Turnitin AI face frequent false positives, pushing students and writers toward humanizers or deliberately clumsier prose.

The controversy is reshaping how publishers, literary prizes, and universities think about authorship and detection.