Skip to content

INDUSTRY FIELD GUIDE

Book edition and excerpt provenance data for AI

A revised edition can change page numbers and substitute an excerpt licensed from another author.

By Fiund · Updated 2026-11-05

NAICS 513130 · Book Publishers

A revised edition can change page numbers and substitute an excerpt licensed from another author. A model citing a passage needs the edition as well as the source. Library of Congress format materials provide publication context. Use original practice text to demonstrate how the record works. Library of Congress recommended formats · Copyright Office recordation overview

Write two short invented editions with a moved section and a replacement passage. Label edition, format, section anchor, author and permission status. Ask which text supports an answer and whether its use is within the exercise’s terms. Include a quote from an excluded dummy source so the rights test is independent of retrieval.

Do not share real manuscripts, author agreements, advance copies, licensed illustrations or classroom materials without clearance. Check edition and use rights separately from ownership of a printed copy.

Describe a book excerpt inventory

Original Fiund exercise

Retrieve the edition-specific passage before checking the stated use

An author creates two short fictional book editions and a dummy excluded-source passage. All text is independently written for the evaluation.

Fictional records supplied to the evaluator
RecordSourceWhat it records
BE1Original edition ASection Harbor on page 4 says The gate opens at dawn. Author alias Writer-A; practice use U permitted.
BE2Original edition BHarbor moves to page 7 and says The gate opens at noon. Same author; U permitted for this passage.
BE3Dummy inserted excerptPage 4 of B contains unrelated Moon passage from Source-X; exercise register explicitly excludes Source-X from U.

Question for the model

What time does edition B's Harbor passage state, and does a correct retrieval of the page-4 excerpt establish permission to use it?

Show the reference answer and checks

BE2 locates Harbor in edition B at page 7 and states noon. BE1's dawn answer belongs to edition A, so page numbering alone would retrieve the wrong edition or section. BE3's page-4 Moon excerpt is explicitly excluded from practice use U. Being able to retrieve or quote the text does not override the supplied exclusion. Score content retrieval and permission matching as separate dimensions.

  • Answering dawn from the superseded edition or page-number match.
  • Treating successful retrieval as evidence that an excluded source is cleared for the requested use.

Fields to preserve

Edition, format and section anchor
Keeps location changes from altering source identity.
Author and excerpt source
Tracks a substituted passage independently.
Use-specific permission status
Separates access to text from the exercise's allowed use.

Test another case

Create an ebook edition with no page numbers but the same Harbor anchor. Test whether the model follows the section and edition rather than inventing a print-page correspondence.

Use original practice text only. Include no manuscripts, advance copies, author agreements, licensed illustrations or protected classroom material.

Download this fictional example and answer key (JSON)

Discuss a contribution

Describe the sources you control, the task they could support and what has already been checked. Use the existing-data assessment for an archive or the collection brief for a new recording project. The contributor agreement defines Fiund’s licensing authority and compensation. A guide or fictional exercise does not establish an accepted collection or a sale.

Primary sources and further reading