Nieman Lab reporter Andrew Deck found that ChatGPT is adding the signatures of real New Yorker cartoonists to fake cartoons it generates, and some of these images have gone viral. More than 15 current and former cartoonists were identified, including Brendan Loper, Harry Bliss and Emily Flake. Loper's pen name, BLOPER, appeared on a viral cartoon that ChatGPT produced and a Facebook user first posted. One tweet of it drew 25,000 likes. OpenAI called the behavior unintended. After Nieman Lab notified the company, ChatGPT began returning a guardrails warning for New Yorker-style requests. A secondary summary says OpenAI's 2024 Condé Nast licensing deal did not cover training on cartoons. The story reached 461 points and 336 comments on Hacker News. For anyone building on image models, a signature reproduced on a generated image is false attribution, not only style imitation. Founders shipping image generation should consider output filters for signatures and watermarks, and should expect rights-holders to cite documented examples like these.