Tech of the Week #4: “Alexa, is AI-Powered Voice Transcription Worth My Time?”

Real-time, AI-powered technology is fast becoming available that takes what we say and transcribes it with an accuracy of up to 95%. It’s being used by everyone, from paediatricians to the media employees that write up interviews as they happen. But Otter is not your typical voice transcription technology.


Alexa, Cortana, Siri

They’re your Tom, Dick and Harry of voice-powered technologies and they have revolutionised the way we go about our day. Alexa, especially, has quite literally become a household name. When she’s not laughing creepily at you as you go about your business, she can do anything from ordering you your next batch of goodies to integrating with smart home technology. In fact, Amazon (who recently debuted their “queue-less”, cashless store named Amazon Go), are apparently working on having Alexa translate in real-time.

Studies, including one by ComScore, say that 50% of all searches will be conducted via voice by 2020. Cortana was built into Windows 10 and the Xbox One by Microsoft to allow just this. This is especially useful on the Xbox. Instead of using the joystick and A button to scroll through an onscreen keyboard, speaking directly at the machine saves plenty of time. It also allows for more natural searches. When typing into Google, we use keywords or short phrases to limit results, but artificial intelligence can deal with an entirely written (or worded in this case) question. That’s because it learns based on what other users are saying when searching for an answer, and understands the context to provide a more accurate result. Also known as machine learning, in a somewhat dumbed down way.


Otter’s Impact on Our Lives

[clickToTweet tweet=”Otter is a voice #transcription #app that takes what’s being said and writes it down for you, like any app that transcribes. The difference here is that this one does it in real-time. #AI #ML #IoT” quote=”Otter is a voice transcription app that takes what’s being said and writes it down for you, like any app that transcribes. The difference here is that this one does it in real-time.”]

Otter is a voice transcription app that takes what’s being said and writes it down for you, like any app that transcribes. The difference here is that this one does it in real-time. Rather than recording what is said and presenting it to you later as traditional services offer, Otter’s transcription is given almost immediately.

The developers behind this machine-learning powered technology are former Google architect, Sam Liang, and Nuance and Facebook employees. Nuance is an award-winning provider of speech and imaging software. They have come together and formed AISense. It comes as no surprise then that Otter can transcribe with accuracy up to 95%. One of the biggest causes of inaccuracy arising from its use of punctuation, or lack thereof, and its lack of ability to properly function in a crowded place.

The app also provides keywords based on what has been said, which could be useful when searching through a substantial transcription to find a particular section. Or maybe, in the case of writing an article, giving a reader a synopsis of what it’s about.


What settings or scenarios could transcription prove useful for?

A hospital is one. During an operation, the surgeon or other members of staff may take notes on what is happening. This could be the amount of one particular drug that has been administered, or a rough idea of what steps were taken. These notes would usually be quite basic, just giving the basics for the most part. Otter would remove the need for this by allowing full details to be transmitted as easy as it is to speak the words. If this was combined with the internet of things, these notes could be transferred between departments faster, potentially saving lives. Or it could be shown to hospital bosses (data protection aside) for use in monitoring the staff and care of its patients.

An office meeting or lecture is another excellent example. The lecture hall can be a noisy place, sometimes drowning the voice of the lecturer, or at least making particular words get lost. Luckily, AI isn’t as quickly thrown off, meaning that (particularly in the case of Otter) by learning different voices, it can pick out the one you need and fill in the gaps. The lecturer could even record his session to reflect on afterwards and improve for future classes, thus giving people an even better chance to learn. In the setting of an office meeting the recording could then form part of the minutes, or as a base point for all employees to refer to.

I, personally, have Ideas come into my head, and they go before I can even find a pen that works. By using a voice transcription service as intelligent as Otter, I could get that dinner recipe or business idea down and ready to refer to when I need it most.

The possibilities are definitely broad, with more becoming available every day that the AI, ML and deep learning technologies advance. The app is currently free and available for Apple and Android phones.


Next in the series, Wednesday 21st March 2018…

Tech of the Week #5: 3D printed houses that are fresh off the press

You can buy a lot of things with $10,000; a family trip to Florida, a 1970s Volkswagen Beetle, and the wedding every girl dreams of. Usually, a house is not an option to buy for so little, especially not a 3D printed home that can be built in 12 to 24 hours. However, Austin-based startup ICON is making this idea a reality…


If you want to talk tech (and have it transcribed), and for more great content, follow me on Twitter @JoshuaOsborn16.


+ posts

Meet Stella


Related articles

Data is the key to unlocking investment for emerging markets

It should be clear then that data, and transport data in particular, is crucial to securing the kinds of public and private investment that will take emerging market economies from recovery, towards sustained (and sustainable) growth.

A New Journey to the Cloud

ERP implementation has changed. And for those companies facing the 2027 maintenance deadline for SAP ECC 6, that is good news. In today’s cloud-first, ‘adopt not adapt model, there are no more white boards.

How to add AI to your cybersecurity toolkit 

A successful implementation of AI in cyber defense requires a solid data governance system, reimagined incident response frameworks, sufficient talent and expertise to manage the new system, and established documentation practices.

The Metaverse: Virtually a reality?

Metaverses have the potential to enable virtual worlds to expand beyond the gaming genre to encompass all manner of social and commercial activities.

Cybersecurity and Cloud: A Look Back at 2022 and What to Expect in 2023

Businesses are continuously reassessing their resources and options to fill their tech stack. In this competitive digital landscape, the innovative use of technology will be something that would generate a competitive advantage for organisations.

Subscribe to our Newsletter