ChatGPT Forges Signatures of New Yorker Cartoonists on AI-Generated Cartoons

ChatGPT Caught Forging New Yorker Cartoonists' Signatures

ChatGPT has been producing fake New Yorker cartoons complete with forged signatures of real cartoonists, creating confusion and raising serious ethical concerns around AI-generated art. The issue came to light after a viral tribute cartoon featuring Dolly Parton and Tim Curry entering heaven together carried the signature "BLOPER"—the pen name of New Yorker cartoonist Brendan Loper—despite him having no involvement in creating it1

1

. The AI-generated comics were created by ChatGPT users requesting "New Yorker-style" cartoons, with the chatbot automatically adding signatures of real artists' names to lend authenticity to the outputs2

2

.

Real Artists Face Misattribution and Personal Branding Violations

Loper told Nieman Lab the experience felt "very surreal," with messages flooding in from his twin brother, friends, and strangers asking if the cartoon was his work3

3

. "My name is my name. It felt very much like a violation of my personhood," Loper explained, emphasizing that his signature serves as his certificate of authenticity and personal branding3

3

. The investigation by Nieman Lab documented more than 15 instances where ChatGPT reproduced signatures of New Yorker cartoonists including Harry Bliss, Emily Flake, Joe Dator, Pat Byrnes, Peter Vey, and Jason Adam Katzenstein1

1

.

Copyright Infringement and Trademark Issues Emerge

The forged signatures add a complex legal dimension beyond typical copyright infringement concerns. Emily Flake, another affected contributor, noted she could "see the DNA" of her work in the AI outputs but emphasized they weren't her creations2

2

. Legal experts point to parallels with Getty Images' lawsuit against Stability AI, where a judge found trademark infringement when AI recreated Getty's watermark3

3

. The signatures represent not just artistic style but intellectual property that identifies the creator's brand.

OpenAI's Licensing Deal Doesn't Cover Cartoon Training

The findings create an awkward situation for OpenAI given its ties to The New Yorker's parent company. In 2024, Condé Nast signed a multi-year licensing deal with OpenAI covering content from brands including The New Yorker, Vogue, GQ, Vanity Fair, and Bon Appétit1

1

. However, a New Yorker spokesperson clarified the licensing deal never granted OpenAI permission to train models on cartoons or use the magazine's logo and art in image generation2

2

. Contributing cartoonists retain copyright of their work under standard freelance contracts, which don't allow licensing for AI training data1

1

.

How AI Training Data Captures Signatures

The New Yorker signatures likely result from widespread internet scraping that siphons content for model training. The ubiquity and popularity of New Yorker cartoons across the internet made them prime targets for inclusion in AI training datasets2

2

. OpenAI declined to explain how its models ingested the signatures, though the mechanism appears straightforward—the AI learned to associate specific artistic styles with particular signatures through exposure to countless authentic cartoons during training1

1

.

OpenAI Implements Guardrails After Investigation

Following Nieman Lab's notification, ChatGPT began displaying a guardrails warning for requests to generate New Yorker-style images: "This prompt may violate our guardrails concerning similarity to third-party content"2

2

. However, as of publication, the chatbot continued signing some generic cartoons with names of real New Yorker cartoonists1

1

. An OpenAI spokesperson stated: "We believe the future of creativity is one that is fundamentally human, and our focus is on building tools that empower human creativity and creators"2

2

.

Broader Implications for Publishers and Congress

Many publishers, including Condé Nast, now push Congress to enact federal laws preventing unlicensed bot scraping2

2

. These same publishers have signed deals with AI companies to recoup revenue lost from previously unlicensed data grabs, creating a complex landscape where past training may have violated terms that current partnerships attempt to legitimize. Watch for potential legal challenges from individual cartoonists whose signatures were reproduced without consent, as personal branding and trademark issues could provide stronger legal grounds than copyright claims alone. The incident underscores ongoing tensions between AI companies' need for training data and artists' rights to control how their work and identity are used.