ChatGPT generates images of fake cartoons in the style of the New Yorker that include forged signatures of real cartoonists like Brendan Loper, Harry Bliss, and Emily Flake. These AI-generated images, often shared on social media, mimic the magazine's style but are not authorized. Loper reported seeing his pen name 'BLOPER' on cartoons he did not draw, calling it a violation of his personhood.
He noted that signatures serve as watermarks and certificates of authenticity in an era of unlicensed sharing. Dozens of New Yorker cartoonists have had their signatures reproduced by ChatGPT without permission or compensation. After being notified, OpenAI added a warning message for such prompts, but the model still occasionally outputs cartoons with forged signatures.
The issue highlights ongoing concerns about AI training data and the misuse of creators’ identities in generated content.
المصدر: Hacker News · لخّصه HeadlinesBriefing