PHOTOGRAPHY VERTICAL

A single photograph carries two rights systems at once: the photographer’s copyright in the image, and the depicted person’s identity and likeness rights. Generative AI sharpens both — archives become training corpora, and synthetic images blur the line between captured and generated. Photography is distinctive in layering two ownership questions in one asset. Copyright generally subsists in an original photograph and belongs (absent agreement otherwise) to its author; separately, the person depicted may hold identity or likeness rights, and commercial use is typically governed by a release. A declaration that speaks only to copyright misses half the picture.

NILP handles the depicted person’s identity and likeness. The training-ingestion / TDM opt-out fields let an archive declare its position on AI training. Source-material records derivation where an image draws on prior work. Output-provenance can distinguish captured from synthetic. The Core Data Record carries asset identity and lineage, and output-licence the permitted uses. What is missing is photography-specific structure tying these together.

1 — Subject-consent / model-release declaration (QUEUED): a machine-readable record of the depicted person’s consent and its scope (editorial, commercial, AI use). 2 — Editorial-vs-commercial use scope (QUEUED): a way to declare the use boundary that releases and many image licences turn on. 3 — Archive & stock licensing inheritance (QUEUED): how a per-image position inherits from an archive- or collection-level licence. 4 — Capture-authenticity signal (QUEUED): a provenance declaration distinguishing a captured photograph from a synthetic or materially AI-altered one.

Legal references on this page are to well-established general doctrines only — no specific statutes, section numbers, dates, or figures are asserted. Qualified counsel should review before any of this is treated as authoritative. The field families listed above are not yet specified.

NILP Architecture for subject likeness. Source Material Architecture (v3.34) for derivation. Output-Provenance Architecture (v3.17) for capture-vs-synthetic. cip.md Generator for operator declarations.