ChatGPT Is Forging Cartoonists' Signatures — and the Problem Runs Deeper Than a Bug
How this was made Verified AI
Every Intellegix briefing is generated from that day's broadcast and run through automated checks before it publishes — with a human paged on any flag. Here is the trail for this edition.
A report from Nieman Lab describes users prompting ChatGPT to generate New Yorker-style cartoon images, with the model producing output that includes what appear to be authentic signatures from specific named living cartoonists. The model has learned to associate the visual gestalt of New Yorker cartoons — including the handwritten name in the corner — with the genre itself, and reproduces that signature pattern when prompted for the style. It is not attempting to forge anything; it lacks a clean conceptual distinction between visual style and specific identifying mark.
The legal terrain is genuinely unsettled. Signature reproduction does not map straightforwardly onto copyright, which governs creative works rather than identifying marks. But depending on jurisdiction, the behavior potentially implicates personality rights, the right of publicity, and passing-off or false-designation-of-origin provisions under trademark law. An image bearing a real cartoonist's name presented as authentic work creates a defamation and fraud vector that is legally serious.
The deeper concern, as framed in commentary on the report, is structural. The model was built by ingesting these artists' work without compensation or consent, and the signature reproduction is the most visible symptom of a wholesale appropriation problem that abstract copyright arguments have failed to make viscerally legible. OpenAI will likely patch the signature reproduction; the underlying question of what models built on artists' creative work owe those artists remains legally unresolved and is not going away.