Control, Provenance, Disclosure, and Access
The September 19 Publishing Standards Brief turns four fast-moving developments into four durable operating records every modern publisher should maintain.

Publishing standards are becoming operational records, not just production guidelines. Four developments make that especially clear this month.

Google has rolled out worldwide controls and reporting for publisher participation in its generative Search features. The Anthropic settlement process has exposed the practical value of clean rights records. EU AI transparency obligations became applicable on August 2. And W3C published a new Candidate Recommendation Draft of EPUB Accessibility 1.2 on September 12.

Together, they point to one practical conclusion: modern publishers need four connected records covering discovery, rights, AI use, and accessibility.

1. Google AI Search Publisher Controls

A Search Console switch is not an SEO strategy. The new global control gives publishers a measurable governance choice: participate, test, or opt out of covered generative Search features. Our proposed AI Discovery Register records the choice, owner, baseline, evidence, and next review date.

Read the full Google AI Search standard →

2. The AI-Era Rights-Provenance Standard

The Anthropic claims process shows why ISBNs, copyright registrations, contracts, reversions, permissions, and source files must be connected. An ISBN identifies an edition. It does not replace registration or prove the current chain of title.

Why this matters to independent publishers: self-publishing often simplifies ownership, but the author still needs evidence for every edition, contributor, licensed asset, and later rights deal.

Build the Rights-Provenance Minimum Record →

3. The Publisher AI Disclosure Standard

“AI was used” is too vague. Publishers need a classification that distinguishes assistive use from material AI contribution and predominantly synthetic work. The record should say what the tool did, who checked it, what was revised, and where disclosure appears.

Use the four-level AI disclosure framework →

4. EPUB Accessibility 1.2 Release Gate

W3C’s September 12 draft makes the operational point unmistakable: accessible content also needs accurate discoverability metadata, evidence, and reader support. An automated validation report alone is not a complete accessibility claim.

Why this matters to independent publishers: accessibility metadata can help readers, retailers, libraries, and procurement systems understand whether a book fits a reader’s needs before purchase or download.

Use the EPUB Accessibility Release Gate →

The Four Records Every Modern Publisher Needs

1. AI Discovery Register for generative-search participation.
2. Rights-Provenance Minimum Record for ownership, registrations, permissions, and source history.
3. AI Use and Disclosure Record for consequential AI use and human accountability.
4. EPUB Accessibility Release Gate for file quality, metadata, evidence, and reader support.

These records do not predict what every court, regulator, platform, or standards body will do next. They give publishers something more durable: a repeatable way to make decisions, preserve evidence, assign responsibility, and update practice when the rules change.

Use this issue as an operating checklist

Do not read the four articles and file them away. Assign an owner for each record, create the template, audit one current title or property, and set a review date. The advantage comes from turning standards into routine operations.

Primary call to action: Start with the AI Discovery Register and build the four-record publishing standard →