An identifier for people is not a test for AI origin

The system

The International Standard Name Identifier (ISNI) is an ISO-certified identifier, ISO 27729, assigned to contributors to creative works, from writers and performers to publishers and producers. It is governed by the ISNI International Agency (ISNI-IA), registered in London, whose own site states its mission is 'to resolve the problem of name ambiguity in search and discovery' by assigning a persistent identifying number and making it available across the global supply chain. ISNI sits alongside other identifiers in the same family, including ISRC for recordings and ISWC for compositions, per the agency's own description. The agency reports, as of retrieval, more than 16.6 million identity records, drawn from over 100 contributing sources.

What the documents establish

ISNI's own technical documentation states that identifiers are assigned through batch loads, online forms, or an assignment-request API, with matches confirmed either through confident matching against existing records or because a name is unique, and that a shared Quality Team at the Bibliothèque nationale de France and the British Library reviews the resulting data. A February 2026 ISNI press release, announcing new guidance for book-industry authors, states that one benefit of the identifier system is 'greater transparency around authorship by distinguishing human, fictional and AI-generated entities through persistent identifiers.' Read against ISNI's own registration process, this means separate ISNI records can be created for a human creator's name and for a distinctly named fictional or AI-attributed persona, keeping the two from being conflated in metadata, rather than ISNI itself analyzing a work to determine whether it was AI-generated.

Craft and rights

For a musician crediting collaborators on an AI-assisted release, ISNI's practical value is disambiguation: it lets a human contributor's name resolve to one consistent, persistent record across retailers, rights systems and royalty platforms, rather than being confused with a same-named or similarly-named party. What ISNI does not do, based on its own materials, is inspect a piece of audio or text and determine whether it was made by a human or a model. Any flag distinguishing human from AI-generated origin has to live in a separate metadata field or a distinctly registered identity, not in the ISNI number itself, which identifies a named party and nothing about how a given work was produced.

Outcomes and open questions

The February 2026 guidance is aimed specifically at book publishing, developed with a publishing-sector consultation group; whether and how a comparable practice of registering separate identifiers for AI-attributed personas takes hold in music metadata and royalty systems specifically is not established by the documents reviewed here.

• Does an ISNI record identify the actual human contributor, or a separate persona or brand name attached to a release?
• Is an AI-contribution disclosure expected to live in the ISNI record itself, or in a different metadata field entirely?
• Has the relevant rights database actually adopted ISNI, or is it referencing a different identifier scheme?

A persistent identifier for people is a foundation correct crediting can be built on, but it is not itself a test for how a piece of music came to exist.