Nieman Lab documents that ChatGPT's image generator produces fake New Yorker cartoons bearing real cartoonists' signatures — whether OpenAI mitigates or denies the forged-signature failure, and whether rights-holders, cartoonists, or regulators respond, decides if generative misattribution of living artists' work becomes a recognized IP-liability failure for frontier image models.

state: seedheat: lowuncertainty: highconvergesscott: highcontent-provenance generative-ip-liability image-generationOpenAI Nieman LabThe New Yorker

What is this?

Nieman Lab — Harvard Nieman Foundation's journalism-futures publication — published a report on October 5, 2026 (by Andrew Deck) documenting that ChatGPT's image generator produces fake New Yorker cartoons bearing real cartoonists' signatures, i.e. falsely attributing AI-generated images to living artists (Brendan Loper is named in connection with one such cartoon) rather than merely imitating the magazine's house style. The supplied snippets confirm the report itself and surface adjacent Nieman Lab coverage of OpenAI/Microsoft training-data court records, but establish nothing about OpenAI's response, the affected cartoonists' reactions, or any legal or regulatory follow-up — that remains the open question the case is tracking.

Why it matters to Scott

Converges with the provenance/cryptographic-trust position in Scott's canon — that asserted authorship must be cryptographically attested rather than visually claimed — by supplying a dated flagship-model receipt: ChatGPT forging living artists' signatures is the exact failure asserted-mark provenance predicts. It also extends his tracked OpenAI artist-guardrail (Ghibli) position from style imitation to forged attribution of named individuals, and it is the first output-side misattribution episode on a radar whose copyright/liability cases (NYT, Anthropic songs, Doe v. GitHub) are all training-side — the corner his canon already flags as 'generative output attribution copyright liability exposure'.

ip:concept.provenanceip:concept.cryptographic-trustwork:technology.openai-apiwork:project.openaiwork:concept.ai-consulting-practiceradar:concept.ai-provenanceradar:concept.provenanceradar:openai-eu-text-provenanceradar:chatgpt-audio-attachment-provenanceradar:concept.copyrightradar:concept.ai-policy

queries asked of Scott's wikis

content provenance C2PA watermarking signed media position
artist style generation opt-out guardrail OpenAI policy Ghibli
image model memorization training data leakage failure mode
generative output attribution copyright liability exposure
agent knowledge base ingestion of untrusted AI-generated media trust
AI product pattern authenticity signals misinformation provenance

Measured heat

now 0 pts/hpeak 70 pts/hcomments 0/hpeers p14momentum: steady2 platformsage 137h

points/hour across evidence · reading as of 2026-10-12 02:59:37.977291+11:00 · deterministic, not a model opinion

How the heat travelled

10-06 03:36 (minted)	⭐ origin echo-reconstructed	"ChatGPT is adding real cartoonists' signatures to fake New Yorker cartoons" — the report is itself the event, a documented flagship-model s Nieman Lab on blog (echo) · attributed from hn.story.49971846 · published time unknown	—
10-05 22:46	first on hacker news · published · lag ?	ChatGPT is adding real cartoonists' signatures to fake New Yorker cartoons rdmuser	—
10-05 22:46	amplified on hacker news 👑	hn.story.49971846 rdmuser	peak 555 · 339 comments · 100% of case engagement
10-06 00:20	our radar first saw it · lag ?	discovery anchor: hn.story.49971846	—

pace: p91 vs 1247 stories at the 96h mark (now 137h old) — ahead of llama-cpp-prompt-lookup-drafting-42x (1.0x), behind moonshot-claude-routing-allegation (1.0x)

Evidence (2) — ⭐ canonical anchor

source	object	author	score	comments
🟧 hn	ChatGPT is adding real cartoonists' signatures to fake New Yorker cartoons	rdmuser	555	339
🟧 echo.blog ⭐	"ChatGPT is adding real cartoonists' signatures to fake New Yorker cartoons" — the report is itself the event, a documented flagship-model s	Nieman Lab	—	—

Interpretation history

2026-10-08T05:42:58Z
HN discussion has cooled (0 pts/hr, 15th percentile); no OpenAI response, artist response, or legal action yet. Case remains a documented flagship failure awaiting the response that would corroborate or escalate it.
2026-10-06T03:44:18Z
grounded: converges/high — Converges with the provenance/cryptographic-trust position in Scott's canon — that asserted authorship must be cryptographically attested rather than visually c
2026-10-06T03:36:18Z
case created — Front-page HN spread on a concrete flagship-model provenance failure with no open case covering generative signature forgery, and an OpenAI or rights-holder response is likely to move it fast.

Decision trace

10-08 18:04attention_routeThe editor compared this story and chose to keep watching.
10-08 16:47attention_routeDocumented provenance failure on flagship model converges with Scott's cryptographic-trust thesis; can wait for briefing in ~1.25 hours as no immediate response or escalation.
10-08 16:42attention_candidatecoverage review: magnitude valve eligible (multi-platform, top-decile engagement); not yet communicated
10-08 16:42repriceHN discussion has cooled (0 pts/hr, 15th percentile); no OpenAI response, artist response, or legal action yet. Case remains a documented flagship failure awaiting the response that would corroborate
10-06 22:22sensor_dirtyvelocity_spike
10-06 16:21sensor_dirtyvelocity_spike
10-06 16:21sensor_dirtycomment_update
10-06 14:44groundConverges with the provenance/cryptographic-trust position in Scott's canon — that asserted authorship must be cryptographically attested rather than visually claimed — by supplying a dated flags
10-06 14:36createFront-page HN spread on a concrete flagship-model provenance failure with no open case covering generative signature forgery, and an OpenAI or rights-holder response is likely to move it fast.