ChatGPT is still churning out New Yorker style cartoons complete with forged signatures from real staff artists.
According to Nieman Lab, viral outputs shared on social media show the chatbot attaching the signatures of more than a dozen professional New Yorker cartoonists to its AI-generated images, often rendered in the magazine's distinctive house style. None of the artists were asked or paid. Cartoonist Emily Flake told Nieman Lab the results felt like "seeing a distant cousin" of her real work. OpenAI now shows a warning when users prompt for New Yorker style images, but the system still produces them, signatures included.
The awkward part is that Conde Nast, the New Yorker's parent company, already struck a licensing deal with OpenAI in 2024 covering brands like Vogue and GQ. A New Yorker spokesperson says that deal never covered training on cartoons or using the magazine's art, and cartoonists keep their own copyrights regardless. So the problem here isn't a missing contract, it's a model that can't tell the difference between what it's licensed to use and what it scraped anyway.
This fits a pattern: publishers cut licensing deals, then discover the fine print didn't stop the scraping that made the deal necessary in the first place. It is also why Conde Nast and others are now lobbying Congress for federal rules against unlicensed bot scraping, rather than relying on company-by-company contracts. A warning label that still lets you draw the forgery is not a fix, it's a disclaimer.