An investigative report by Nieman Lab revealed that OpenAI’s ChatGPT chatbot produced cartoons mimicking The New Yorker’s style and appended them with original signatures of its real artists. This action sparked widespread outrage among creators amid legal questions surrounding intellectual property infringement.
- Investigative report uncovers forged artist signatures
- Spread of fake illustrations and reactions from affected artists
- Licensing agreement loopholes and protected copyright
- Formal protection measures and the legal outlook on the impasse
- Frequently asked questions
Investigative report uncovers forged artist signatures
A comprehensive investigative journalism report published on Monday, October 5 by Nieman Lab, an organization specializing in journalism and media affairs, showed that the popular chatbot ChatGPT, developed by OpenAI, generates cartoons that imitate the unique artistic style of the renowned American magazine The New Yorker. The system signs them with the names and signatures of real artists and illustrators who never participated in drawing or reviewing them. Investigative journalist Andrew Deck tracked the signatures and pseudonyms of over 15 well-known illustrators and contributors appearing clearly on AI-generated images without obtaining their consent or paying them any financial compensation.
This serious incident goes beyond artistic imitation or general style adaptation, reaching the point of presenting automated outputs as official work published by a specifically named artist, which clearly misleads the viewing public. Although OpenAI added new warning messages after being contacted by the investigative team, technical checks confirmed that the software system continued to add real artists’ signatures to a number of subsequently generated illustrations.
Spread of fake illustrations and reactions from affected artists
The investigation focused on the experience of cartoonist Brendan Loper, known for signing his work with the shorthand name “Bleubird.” His artistic signature appeared on an AI-generated illustration created by a Facebook user who asked the chatbot for “a New Yorker-style cartoon” following rumors surrounding the deaths of entertainers Dolly Parton and Tim Curry in late August. The image went viral on social networks, and many viewers believed it was created by Loper himself, who expressed his shock by saying: “I am not a person looking for territorial disputes, but my name remains my name and no one has the right to impersonate it.”
Loper was not the only victim, as the investigation found names and signatures of elite magazine artists such as Emily Flake, Harry Bliss, Joe Dator, Pat Byrnes, George Booth, and Liza Donnelly. Flake commented on the illustrations by saying that she sees the genetic footprint of her drawing style, but feels as though she is looking at a strange creature that resembles a distant relative and has nothing to do with her actual creative work. Meanwhile, Jason Adam Katzenstein, who has published his cartoons in the magazine for 12 years, described the outputs as “a distorted and unsettling simulation of an impression of a joke,” warning that these software programs seek to marginalize human creators and replace them with low-cost algorithms.
Licensing agreement loopholes and protected copyright
These findings raised highly complex legal questions, especially since OpenAI signed a historic content licensing and usage agreement in 2024 with Condé Nast, the parent company of The New Yorker. However, a magazine spokesperson firmly stated that the terms of that agreement do not grant the developer any right to train its models on cartoons, nor do they permit it to exploit the magazine’s logo or artistic designs in generating AI images.
The spokesperson stressed that cartoonists retain full intellectual property and independent copyright for their artwork, which was confirmed by a review of standard contracts the magazine concludes with contributors that explicitly prohibit licensing drawings for training artificial intelligence algorithms. In response, an OpenAI spokesperson merely stated that the company believes the future of creativity will remain essentially human, expressing appreciation for community feedback on technical errors without providing any clarification on how the models learned to mimic those signatures with striking accuracy.
Formal protection measures and the legal outlook on the impasse
Following the exposure, ChatGPT’s generation interface displayed a warning message for certain prompts related to New Yorker cartoons, indicating that the request “may violate safety guidelines regarding third-party similarity.” However, the system still generates the images and sometimes ignores the restriction, appending the signatures anyway. Professor James Grimmelmann, a professor of law at Cornell University, commented that any lawsuit arising from this practice will focus primarily on whether the algorithms reproduced substantial protected portions of the artists’ work, rather than focusing solely on drawing the signature by itself.
This dispute opens a new chapter in the raging legal battle between creator unions and major AI companies, as artists view the use of their names to falsify artwork without compensation as a blatant assault on professional identity and moral rights protected by international covenants, calling for urgent regulatory intervention to halt these violations and control the performance of algorithms on public platforms.
Frequently asked questions
Question: What did the Nieman Lab investigation reveal about ChatGPT?
Answer: The investigation revealed that the platform generates cartoons mimicking The New Yorker’s style and appends them with forged signatures of more than 15 artists without their permission.
Question: Does OpenAI’s agreement with Condé Nast cover training models on cartoons?
Answer: No, the magazine confirmed that the agreement does not include licensing cartoons or using them to train artificial intelligence, as artists retain their independent intellectual property rights.
Question: How did OpenAI respond to these accusations?
Answer: The company added a warning regarding guideline violations, but did not clarify the mechanism by which the model acquired the artists’ signatures, limiting its response to emphasizing support for human creativity.
