The phrase 'this looks like AI' has mutated from observation into indictment. A Verge piece by Jess Weatherbed tracks the emerging ecosystem of 'human-made' badges and certifications, a kind of artisanal provenance label for creative work. Meanwhile, Fast Company reports that a New York Times critic used AI to write a book review, and the discourse detonated. The question underneath both stories is the same: can authenticity be verified, or only performed?
The Certification Trap and Creative Provenance
What's happening now mirrors something the art world has wrestled with for centuries. A new study on Dalí's decaying 'Temptation of Saint Anthony' found that the artist's own varnish choices are destroying the painting from within. Authenticity, it turns out, can be self-undermining. The material you use to assert permanence becomes the vector of decay. The AI-free badge risks the same irony: the more aggressively you certify humanity, the more you signal anxiety about it. Robert Rauschenberg, as Hilton Als writes in The New Yorker, spent his career bending photography to make the real feel realer. He didn't certify. He transformed. That gap, between transformation and certification, is where the current debate lives.
What Criticism Actually Requires
The NYT AI review scandal isn't really about cheating. It's about what criticism is for. Good criticism is the record of a sensibility encountering a work. You cannot outsource a sensibility. You can outsource summarization, comparison, even argument structure. But the thing that makes a review worth reading is the irreducibly personal residue of someone actually having experienced something. That's not a romantic claim. It's a functional one. If the AI-free certification movement wants to mean anything, it needs to articulate what human attention produces that machine attention cannot, and that argument has to be made through the work itself, not a badge in the corner.