Kvíz spojený s průvodcem · Střední úroveň

AI for Oral History Transcription Quiz

Test AI oral-history transcription: draft versus source audio, diarization, dialect, narrator consent, transcript review, preservation, and access

Související vodicí cestyAI For Oral History Transcription
Otázka 1 z 8

When an AI transcript conflicts with the recording, which source should guide correction?

Náhled každé otázky8 otázky

Odpovědi a vysvětlení se zobrazují, jak budete v kvízu.

  1. When an AI transcript conflicts with the recording, which source should guide correction?
    • The model’s most fluent sentence
    • A summary generated from the transcript
    • The original audio recording
    • A guess based on the interview topic
  2. Why should a narrator be offered transcript review when possible?
    • It guarantees the speech model is perfect
    • It replaces the need to preserve audio
    • It permits the editor to rewrite the interview freely
    • They can correct names, context, privacy choices, and how their words are represented
  3. What can language-model context do when audio is unclear?
    • Produce a plausible completion that was not actually spoken
    • Guarantee the exact words
    • Remove all speaker overlap
    • Preserve every pause
  4. How should an editor handle an inaudible passage?
    • Insert the most likely phrase without notation
    • Mark uncertainty rather than silently guessing from context
    • Delete the audio segment
    • Ask the model to complete it confidently
  5. What should be checked before sending a restricted interview to a cloud transcription service?
    • Whether the tool can export time-stamped text, without checking how it handles uploaded audio
    • Whether the service appears on an approved list, without reviewing its retention or training terms
    • Consent terms and vendor retention, access, training, and deletion practices
    • Whether the vendor uses a modern website design
  6. Why preserve the original recording alongside corrected transcripts?
    • The transcript is always more authoritative
    • The audio can be discarded after transcription
    • The recording proves every memory is historically accurate
    • The recording lets future users verify wording and interpretation
  7. Speaker diarization can fail in which common way?
    • Overlapping voices or short turns can be assigned to the wrong speaker
    • Diarization changes the recording’s sample rate
    • It proves the identity of each voice
    • It always preserves pauses accurately
  8. What does OHA’s guidance emphasize about interview use and release?
    • Automatic public release after transcription
    • Informed consent, transparency, and clear parameters for preservation and access
    • A ban on all technology use
    • A transcript that omits narrator preferences