ChatGPT Found Replicating Real New Yorker Cartoonists' Signatures on AI-Generated Images

False Attribution and Viral Misinformation

ChatGPT’s image generation model has been found to replicate the actual signatures of New Yorker cartoonists on AI-generated fake cartoons, creating false attribution rather than just style imitation. A prominent example involved a viral image of Dolly Parton and Tim Curry in heaven, which received 25,000 likes on Twitter and bore the signature BLOPER of cartoonist Brendan Loper, who did not create the work. Loper discovered the misuse via Instagram DMs in May 2026 and views the replication of his signature as a violation of his personhood and brand.

Scope of Affected Artists

Investigative testing by Nieman Lab identified signatures of more than 15 New Yorker cartoonists used without permission or compensation. Confirmed artists include:

• Harry Bliss
• Emily Flake
• Joe Dator
• Pat Byrnes
• Peter Vey
• Jason Adam Katzenstein
• George Booth
• Liza Donnelly
• Ellis Rosen
• Saul Steinberg
• Warren Miller

While the model mimics the magazine’s 101-year-old house style (black-and-white, single panel, concise backgrounds), some punchlines remain unclear, and not all generated images contain signatures.

Legal and Contractual Context

Condé Nast, the publisher of The New Yorker, signed a multi-year licensing deal with OpenAI in 2024 but explicitly stated it never granted permission for cartoonists' work to be used for LLM training. The replication of signatures and the New Yorker logo alongside cartoons is prohibited under these agreements. OpenAI declined to explain how its models learned these signatures but acknowledged the issue as a bug, stating they would address it. Following notification, ChatGPT began displaying guardrail warnings for prompts requesting New Yorker-style content, though some generated images still included real artist names.

Artist Reactions and Legal Implications

Cartoonists describe the signature replication as false impersonation and a deeper violation than copyright infringement. Jason Adam Katzenstein noted that attaching his name to work he did not support is akin to putting words in his mouth. Joe Dator compared the intrusion to credit card hacking, stating, "Credit card hacking didn't copy me, but this does." Pat Byrnes provided evidence that over 12 ChatGPT-generated cartoons bore his signature ending with a period, a specific stylistic trait he has used since his teenage years.

Legal experts like James Grimmelmann from Cornell Law School suggest that while copyright claims may be weak due to the distinction between style imitation and direct copying, right of publicity lawsuits could offer a path for artists if commercial use is proven. Currently, there is no evidence of AI-generated cartoons being sold, but signs indicate AI is beginning to replace professional cartoonists' work, as seen in the Baltimore Sun's decision to hire AI for its opinion section and lay off Pulitzer finalist Kevin Kallaugher.