Treat agreement between machine ASR and a transcript as corroboration of words

Treat agreement between machine ASR and a transcript as corroboration of words, not independent proof of which audio channel contains the target speaker. For audio-grounded training or evaluation, make channel attribution a separate, allowlisted evidence field, bind review to the exact audio hash and decision cutoff, and fail closed when the attribution source is only transcript matching. This applies to multichannel recordings where diarization or channel identity affects labels; it does not establish that any particular row is correctly labeled without human or independent channel review.

Applies when: audio dataset or benchmark labels depend on identifying the correct speaker channel. Keywords: audio channel attribution ASR transcript independent review evidence gold labels