The Pulse
ChatGPT Put New Yorker Cartoonists’ Signatures on AI Images
Nieman Lab found ChatGPT-generated cartoons bearing the signatures of more than 15 New Yorker cartoonists, including Brendan Loper and Emily Flake. OpenAI added a warning about prompts resembling third-party content after the outlet raised
AI.info Team · Written with AI ·
OpenAI says it wants to build tools that empower human creativity; New Yorker cartoonists say ChatGPT has attached their names to images they did not draw. Nieman Lab documented more than 15 of the magazine’s cartoonists whose signatures appeared on AI-generated cartoons, and found some names still turning up after OpenAI added a warning to certain prompts.
The dispute is not only about images that resemble a magazine’s cartoons. A signature tells viewers who made a work, and its appearance on an image can falsely suggest an artist created or approved it. Nieman Lab’s reporting found examples bearing signatures associated with Brendan Loper, Emily Flake, Joe Dator, Pat Byrnes and other contributors.
A viral cartoon carried Loper’s name
One image circulated online in late August, after the deaths of Dolly Parton and Tim Curry. It showed the entertainers at heaven’s gates and carried “BLOPER” in the corner—the pen name of New Yorker cartoonist Brendan Loper. Loper had not drawn or signed it; the image was generated with ChatGPT after a Facebook user asked for a “New Yorker-style cartoon,” Nieman Lab reported.
The image’s circulation brought messages from people asking Loper whether it was his work. He had already heard about an earlier ChatGPT image bearing his name in May and found more examples posted online during the summer. “I’m not a territorial person, but my name is my name,” Loper told Nieman Lab. “It felt very much like a violation of my personhood.”
For Loper, a signature marks authorship, not just style. Brendan Loper, New Yorker cartoonist, told Nieman Lab, “The signature for me is the certificate of authenticity. I only put it on work that I’m putting on the internet that I feel needs to be recognized as from my hand.”
OpenAI added a warning, but signatures persisted
Reporter Andrew Deck tested ChatGPT and found signatures belonging to established and past New Yorker contributors, including George Booth, Liza Donnelly, Ellis Rosen and Saul Steinberg. The images varied: some had no signature, and some used gibberish names. Deck documented more than 15 real cartoonists’ signatures in the examples he examined.
After Deck notified OpenAI, ChatGPT began displaying a warning for some requests to make New Yorker-style cartoons: “This prompt may violate our guardrails concerning similarity to third-party content.” But Nieman Lab said that, as of its October 5, 2026 publication, some generated cartoons still carried real cartoonists’ names.
OpenAI’s response acknowledged reports of unintended model behavior without explaining why the signatures appeared or how the system would address them. OpenAI spokesperson, in a statement to Nieman Lab, said: “We very much appreciate the community flagging bugs and unintended behavior by our models, so we can address them.”
The licensing deal does not cover the cartoons
Condé Nast, The New Yorker’s parent company, signed a multi-year licensing agreement with OpenAI in 2024, but the terms are not public. A New Yorker spokesperson told Nieman Lab that the publisher had not given a large language model developer permission to train on its cartoons. The spokesperson also said the deal did not authorize ChatGPT to reproduce the magazine’s logo.
OpenAI declined to answer Deck’s questions about whether or how its models might have ingested signatures from the magazine’s archive. The origin of the signatures therefore remains unestablished: Nieman Lab said they could have been scraped from the internet, but did not verify that explanation. The outlet also reviewed standard contracts for contributing cartoonists that do not permit licensing their work for AI training; the artists retain copyright, although the magazine has first publication rights to commissioned cartoons.
A false signature raises a different question from style
Emily Flake, a New Yorker contributor since 2008, told Nieman Lab that some generated images echoed parts of her visual style but blended elements from multiple cartoonists. The signature, she said, was unmistakably hers.
Cornell Law School professor James Grimmelmann told Nieman Lab that a signature’s appearance may not settle a copyright claim, which would more likely turn on whether a specific drawing was reproduced. “Attribution is a very small part of the fair use inquiry,” he said. He also said a right-of-publicity claim would require evidence that the images were used commercially; Nieman Lab found no evidence that the cartoons were being bought or sold.
The reporting documents a concrete failure of attribution, but does not establish the training source or resolve what legal claims the artists might have. For now, the clearest measure is the image itself: ChatGPT-generated cartoons that carry real contributors’ signatures, even as the company’s warning appeared and some names continued to show up.