How AI transcription is helping business and education through the pandemic [Q&A]
The COVID-19 pandemic has seen large numbers of people adapting to new working patterns and ways of interacting.
AI-based transcription platform Otter.ai has seen a fivefold increase in demand in recent months and has transcribed 750 million minutes from over 25 million meetings. We spoke to Otter's founder Sam Liang to find out more about how it's helping to shape the future of work and education.
BN: Can you tell us a bit about the Ambient Voice Intelligence technology that underpins Otter?
SL: At Otter.ai we have developed our own proprietary stack, including automated speech recognition, speaker separation (diarization) and identification, word synchronization, and summarization. The Ambient Voice Intelligence technology uses proprietary artificial intelligence to generate secure, shareable, searchable, rich notes in real-time that combine audio, transcription, speaker identification, inline photos, and key phrases. Otter is different from the likes of Alexa or Siri as it is able to process multi-speaker conversations, like a long Zoom meeting.
BN: How do you cope with new words, technical terms, acronyms, etc?
SL: Transcribing text has many challenges to overcome, from new words or acronyms, different accents, how people pronounce certain words, background noise, how close someone is to the microphone, etc.
Otter.ai's system is built from the ground up to handle all the complexity required for transcribing multiple participants. We use millions of hours of recordings to train our deep learning algorithm. We plug in huge amounts of data to train our model, in a very large neural network with many nodes, which creates a huge matrix from thousands of different variations of words, technical terms, acronyms. By using natural language processing (NLP), we are also able to understand the partial meaning of the words. So which words that actually usually go together, or which words rarely go together, to make the transcription model work even more efficiently.
The speech recognition technology itself has a lot of AI and machine learning to separate the noise -- we actually inject in our training system a lot of the noise into the training data to make it harder, so that the model can learn how to handle the noise, as well as accents and new words.
BN: How does the new Live Notes feature differ?
SL: Otter Live Notes enables participants to open a secure, live interactive transcript directly from a video conference, or after a meeting. It is ideal for business users working from home and students attending classes online who want to highlight, share, search, and review their notes. What we all know though is that if you are to 'miss' a meeting, colleagues notes or updates can often be lacking, especially when people are finding remote collaboration more mentally challenging with 'Zoom Fatigue'. Team members can feel confident that their colleagues will generate and share secure quality notes so they can quickly get the information they need when they miss a meeting through our Live Notes function.
Otter Lives Notes has also continued to add new features to become the must-have business and education communications and collaboration product. Over the summer, Otter For Events was introduced to add previously untapped value from webinars and virtual events by capturing conversations and turning them into highly accurate interactive, collaborative live transcripts. More recently Live Video Captioning was introduced to remove any barrier to digital business communications that companies face, boost collaboration, as well as help to meet accessibility needs.
These rich features come with Otter Live Notes now and will now appear within Zoom for Otter for Business and Zoom Pro subscribers or higher.
BN: How can transcription services help with remote learning?
SL: There is a distinct realization that one of the central effects of the pandemic is that it has sped up an immediate and permanent shift in favour of remote learning, creating a new educational experience. Lecturers especially have had to challenge themselves to reconsider fundamental aspects of how their students learn best. This presents serious technology issues as, in reality, the classroom has changed little since the first schools opened in the 17th century. In fact, I would argue that technology has helped to facilitate a 're-imagination' of just about every sector -- except education.
Transcription services address one of the challenges that has arisen with remote learning over video conference tools, the balance of taking notes and engaging with the lecture on the screen, both central pillars of the learning experience. Transcription services that transcribe lecturers in real-time and allows for highlighting of notes during virtual lectures is providing students with a hybrid learning opportunity. The duality enables students that might learn best by reading information rather than hearing it, but also students that just want to focus purely on the lecturer can receive notes that they can refer to after the lecture.
BN: Do you think that increased levels of remote working are here to stay?
SL: I do. Love it or hate it, all signs point to a permanent move toward some form of remote working. The changes we've seen already has created both opportunities and challenges. The advantages of a reduced commute and the ability to control your own work environment has led to cost savings and a more flexible workforce. Managing a disparate workforce and the blurring of work and home life has been the clear downsides for many.
While there may not be the dramatic shift to a 100 percent remote working that some are predicting, we aren’t going back to 9-5 in the office. There will be a hybrid model employed as employees will force companies -- even those that were not the biggest proponents of having a virtual workforce -- to become more flexible with traditional working hours and business practices.
The remote work environment will continue to evolve over time and has and will require new ways of thinking to create the most productive and collaborative workforce. No doubt we're all going to need to continue to adapt and improve our team communication, engagement and productivity wherever possible to make it work.AI