Cartoonist Brendan Loper began receiving messages from people who believed he was the creator of a cartoon that had gone viral online. His pen name appeared on the drawing, but the image had actually been generated using ChatGPT.
اضافة اعلان
Nieman Lab documented the appearance of signatures belonging to more than 15 cartoonists who contribute to The New Yorker on AI-generated cartoons.
These names can give images circulating online a false sense of credibility, making them appear to bear the genuine signatures of their creators, even though the artists had no involvement in producing them.
One example was a cartoon created by ChatGPT that went viral online bearing the signature “Bloper,” the pen name used by The New Yorker cartoonist Brendan Loper, despite the fact that he had neither drawn nor signed the image.
Why does signature imitation raise concerns?
Cartoonist Emily Flake told Nieman Lab that drawings bearing her pen name appeared to combine the styles of several artists, but the signature on them closely resembled her actual signature in a recognizable way.
In 2025, OpenAI stated that ChatGPT could imitate the artistic style of a particular studio but could not replicate the distinctive style of an individual artist.
However, reproducing artists’ signatures raises a different issue concerning the attribution of artistic works to their creators.
The problem extends beyond imitating a drawing style to associating an artist’s name with satirical content they neither wrote nor illustrated.
People who reshare an image may not have access to the original prompts used to generate it through ChatGPT. Meanwhile, artists whose names appear on these images may find themselves having to answer questions about work with which they have no connection.
That is exactly what happened to Brendan Loper.
Have safeguards succeeded in preventing signature imitation?
After Nieman Lab brought the issue to light, ChatGPT began displaying warnings indicating that requests to create cartoons in The New Yorker’s style might conflict with its policies concerning similarities to third-party copyrighted content.
However, the warnings did not resolve the problem, as the publication was still able to generate other cartoons bearing the names of real artists.
OpenAI said it welcomes reports of unintended behavior by its AI models.
The case also recalls a separate issue involving ChatGPT’s image-generation safeguards that emerged in June and persisted even after the company introduced changes to address it.
Researchers had found that making minor changes to text prompts could still allow some restrictions on image generation to be bypassed.
The exact method by which the image-generation tool managed to reproduce cartoonists’ signatures remains unclear.
Condé Nast, the publisher of The New Yorker, signed a content agreement with OpenAI in 2024.
However, the magazine confirmed that it had not licensed its cartoons for use in training AI models.
OpenAI, meanwhile, declined to explain how the names of real artists appeared in images generated by its tools.
How can readers verify a cartoonist’s identity?
OpenAI provides a public image-verification tool that searches for indicators associated with content created using its products.
Finding such indicators may help determine whether an image was generated using the company’s tools, but it does not prove that the artist whose name appears on the cartoon created it or approved its publication.
Conversely, failing to find indicators of AI use does not necessarily mean that an image was created by a human.
Resharing or editing images can remove their metadata or make embedded digital watermarks more difficult to detect.
The effectiveness of verification tools also varies depending on when an image was created and the method used to produce it.
Therefore, if readers come across a cartoon bearing the signature of a well-known artist, they should look for it in the artist’s portfolio or the official archive of the media organization for which they publish before attributing the work to them.
The case highlights a new challenge that AI poses to the artistic and media communities. The issue is no longer limited to the ability of AI tools to imitate creative styles; it also includes the possibility of giving AI-generated works a false identity by using the names and signatures of real artists without their involvement.
