ChatGPT is adding real cartoonists' signatures to fake New Yorker cartoons

Commenters focused on ChatGPT producing New Yorker-style cartoons that include real cartoonists’ signatures, with many treating it as theft and a legal issue. Insimwytim and AnimalMuppet urged lawsuits, MarkusQ called it “Plagiarism as a Service,” and gwern and others described having to erase false signatures from their own AI images. Several commenters framed the problem as impersonation and libel: baubino relayed cartoonists’ feelings that having their names attached to work they don’t endorse is worse than ordinary copyright violation. Practical fixes were proposed too: cm2012 argued AI should default to leaving signatures out, Johnnyanmac and DonsDiscountGas suggested visible watermarks or disclosure, and SchemaLoad and others called for clearer legal remedies.

Opinion split over cause and culpability. Some, like kzsh and saalweachter, treated signatures as statistical artifacts of training data - a predictable cluster of pixels rather than intent - while antonvs and others pointed out differences between image generators and LLMs and disputed blanket claims about “reasoning.” gruez and bulder raised legal and market complexities (satire, trademark, user responsibility), zzzeek and capsadmin framed the issue within broader debates about ethical datasets and worker harm, and critics like drivebyhooting and PunchyHamster tied it to worries about automation, business models, and whether ethically sourced, high‑quality models are even feasible.