HUMAN TRANSCRIBERS V AI

Human transcriptionists, compared to speech-to-text AI (artificial intelligence) transcription software programs, can achieve better, more accurate results because humans:

 

• Cope better with background noise and filter out what’s unnecessary

​​

• Understand different accents and dialects.

​​

• Differentiate easily between homophones (eg: to, too, two; their, they’re, there)

​​

• Can identify several speakers more easily, by the different tone of their voice, the location from which they are speaking (left/right/centre), accents, speed of speech and so on.

​​

• Can pick up unusual turns of phrase, slang, jargon, foreign words, Latin sayings, medical terms, turns of phrase, acronyms, and cope better with mumbling, or incomplete words.

​​

• Realise when proper nouns are being used and research to check the spelling accordingly, as it relates to the subject matter of the audio file.

​​

• For regular customers, we keep a glossary of speakers’ names, organisations to which they may refer, acronyms and so on, to ensure the maximum possible accuracy.

​​

• Are more easily able to produce an intelligent verbatim transcript (a popular style of transcription that represents recorded speech into text while editing out the fillers and repetitions that may distract from getting at the content of the interview.)

​

​

​