A retired civil servant from Trinidad won the Commonwealth Short Story Prize. The institution that awarded it discovered it had never actually thought to ask how the story was written.
The assumption hiding in this crisis is simple. Literary prizes evaluate writing, not writers. The text arrives anonymous or pseudonymous — judges read it, merit decides.
Except nobody actually believes this. Not really. Every serious literary prize has always trafficked in authorship mythology.
The Man Booker judges aren't looking for good sentences alone. They're looking for sentences that prove something about how a consciousness works. This is why a debut memoir from an unknown carries weight that a technically superior ghost-written work does not. We pretend we're evaluating product. We're evaluating origin story.
A handwritten manuscript feels like evidence. A book contract with your name on it suggests you existed before the words did. But the moment verification becomes possible. When a judge can run a text through a detection algorithm, or ask for process notes, or demand to see your drafts — the whole premise collapses. Not because AI changed what makes writing valuable. Because we've suddenly asked institutions to verify a claim they've never actually had to defend.
The real question isn't whether this writer used AI. It's whether literary prizes want to keep pretending they're agnostic about how a story gets made, or whether they're finally willing to name what they've always rewarded. They've always rewarded not just the text, but the evidence of a particular kind of human struggle producing it.
If they choose the second path, the evaluation changes entirely. If they don't, they're admitting they never cared how it was made at all, only that they could plausibly believe they didn't need to ask.
Pick one award or recognition you give (best employee, best friend's advice, most impressive work)—write down whether you're actually evaluating output or your belief about their process, then notice where you've never asked for proof.