Ariel Utnik – Say What? How Verbit is Transcribing at Scale with AI and HI
One of the hardest challenges of any AI is to be able to understand natural language. The first time you spoke at Siri or Alexa, you probably gasped at how amazing it was that the machine could understand you, and almost respond with clarity. Or I’m sure you’ve had the experience of watching a live chat online and laughed at the subtitles.
Seriously though have you ever considered how hard it is for an AI to understand the various languages we have, not to mention accents?
In this podcast I had a chat with the CRO and GM at AI Transcription leader Verbit, Ariel Utnik. We discussed this various challenge, and how through a combination of AI and Human Intelligence (HI) we can, in real-time deliver 99.9% accuracy on transcription.
There were a number of interesting outtakes;
- Why transcription is so important, especially for those that are hard of hearing.
- How transcription might be weaved into live events, how it’s already included in live feeds, and why we are likely to see more of it.
- How challenging and yet how important it is, to ensure accuracy in transcription.
- How the AI learns over time, with the help of human intelligence.