AI Won a Literary Prize. The Judges Couldn't Tell

In May, a conceptual artist posted a Monet painting on X and told everyone it was AI-generated, then asked viewers to explain what gave it away. The replies were brutal: emotionless, no composition, smudgy, high-school-level work.

Every one of the 1,400 critiques was aimed at an image genuinely created by Claude Monet. The label overrode their eyes. People weren't unable to look at art. They were unable to look at art they'd been told a machine had made.

The meaning of a Monet has never lived only in how it looks. It is a physical object made at a specific moment, when Impressionists were trying to paint how the eye sees rather than what it sees.

In 1935, the critic Walter Benjamin called this quality an artwork's "aura": its presence in time and space, its unique material history. He thought mechanical reproduction was eroding it. Generative AI raises a sharper question, because it doesn't reproduce existing work. It manufactures new surfaces that mimic creative expression without the human process behind it.

Two other recent stories suggest something is already breaking. A Nobel laureate admitted to using AI in her writing, and a respected literary prize may have been won by an AI-generated story.

Keeping up with AI shouldn't feel like a second job.

But between the new tools, viral posts, and endless hot takes, most people spend hours every week trying to figure out what actually matters.

The Rundown AI fixes that. 

It's a free newsletter that gives you the AI news, tools, and tutorials you actually need to know. All in just 5 minutes a day.

Over 2M professionals at companies like Apple, Google, and NASA already read it every morning to stay ahead.

Plus, if you complete the quiz after signing up, they'll recommend the best tools, guides, and courses for your specific job and needs.

Sign up to start learning.

The financial damage to creative work is well documented. Researchers at Harvard Business School and Imperial College London tracked 1.3 million freelance postings across 61 countries. Within eight months of ChatGPT's launch, demand for writing fell around 30%, and graphic design dropped 17%.

A February 2026 study from Ramp Economics Lab found freelance spending fell from 0.66% of company budgets in late 2021 to 0.14% by mid-2025. Spending on AI providers like OpenAI and Anthropic rose from zero to nearly 3% over the same period.

More than half of businesses using freelancers in 2022 had stopped entirely by 2025. The middle tier of commercial creative work, from copywriting to logo design, is being hollowed out. That part of the story is straightforward: a cheaper, faster tool displaces human labour.

The assumption was that high culture, literary fiction and fine art, would stay safe, because the value there is supposed to lie in the human process itself. That assumption is now under pressure.

On 12 May 2026, Granta published five regional winners of the Commonwealth Short Story Prize. Within days, the Caribbean winner, "The Serpent in the Grove" by Jamir Nazir, was run through Pangram's detection software and graded 100% AI-generated. Wharton professor Ethan Mollick confirmed the finding independently.

Journalist Vincenzo Barney, writing for UnHerd, found the problem ran deeper. Four of the five winners had little or no online presence, and three stories scored between 92% and 100% for AI. Granta admitted its staff had not read the stories beyond a copyedit.

Detection tools are not flawless, and several universities have banned them over false positives. But a University of Chicago Booth test in August 2025 found Pangram scored near-perfect accuracy on longer passages, with essentially no false positives.

The detection scores alone prove nothing. The pattern does the work: three winners flagged above 90%, four with no literary trail, and a magazine confirming nobody on staff had read what it published. The jury had praised the prose as assured and the characters as utterly believable.

By most accounts the flagged prose was mediocre, full of images that sound meaningful without meaning anything. This wasn't brilliant AI fooling the experts. It was middling writing passing through a system whose standards have drifted toward what AI produces.

Barney traces the writing drift back to around the mid 2000s when writers were forced to play to the rules of algorithms to get attention. Stanford critic Mark McGurl argued that postwar creative-writing programmes produced a more homogeneous, professionalised style of American fiction. His later work described how Amazon's model turned novels into products optimised for customer satisfaction.

An earlier New York Times quiz pointed the same way. Of 86,000 people, 54% preferred AI writing to passages by Cormac McCarthy, Ursula K. Le Guin and Hilary Mantel. Readers described the human work as having edges they didn't expect, and the AI as cohesive, monotone, beige.

The models were trained on this already-flattened output. The result is that AI prose and much contemporary writing have converged on a shared smoothness. The judges couldn't tell because the style they reward, lush but vague and rhythmically predictable, is the style AI generates.

Every creative technology has triggered fears of lost authenticity. When photography arrived, painters predicted the death of their craft. Instead it pushed painting toward Impressionism and abstraction, the things a camera couldn't do.

Eli Lehrer argued in Tech Policy Press that AI expands the range of possible ideas, shifting the real skill toward choosing what is worth pursuing. The science fiction writer Ted Chiang is less convinced, comparing AI-assisted writing to bringing a forklift into the weight room: you never build the strength yourself.

The photography parallel may not hold. A painting and a photograph occupy different perceptual registers, and you can see which is which. AI text and human text sit on the same page, distinguishable only through context and trust, both of which the Commonwealth case shows are degrading.

Benjamin worried that reproduction would strip art of its aura. AI may do something stranger. It floods the temple with convincing copies until the congregation can no longer tell what is sacred. The Monet still hangs in its Paris gallery, unchanged. What has changed is whether we trust ourselves to see it.

Sign up to get a daily bulletin newsletter, an ad free audio version of our videos and regular insights, automations and advanced prompts.