Why virtual assistance technology is the future of the internet
- 11 July, 2017 16:49
Picture: Vanity Mirror by Blinding (Flickr).
Voice of our generation
As the internet of things rapidly becomes the internet of everything, it is becoming clearer that humans will master these devices not through text or touch, but through our most natural and effective channel of communication: our voices. Scientists have been trying to make computers understand human language since the 1950s, but until recently progress had been painfully slow. But in the five years between 2010 and 2015 speech recognition accuracy jumped by around 25% –– more progress than was made in the preceding four decades. With these rapid advancements in automatic speech recognition (ASR) software, it was suddenly a viable feature for commercial products. There are clear benefits to using speech as an input: we can speak faster than we can type; it requires no hands; and as the software remembers more about us it will begin to deliver more relevant results. Most of us are already talking at our phones, but it’s in our homes and workplaces where we could see the biggest changes, with virtual assistants by Amazon, Google, and others all vying for our attention.
Voice tech you can have now
Launched in 2014, the first generation Amazon Echo set the standard for voice-activated smart home technology with its ability to connect to other smart devices, read the latest weather and headlines, manage to-do and shopping lists, as well as stream music. An update in September 2016 added extra features, including the ability to hail an Uber, order a pizza from Domino’s, as well as stream straight from Spotify. Using far-field microphone technology, Echo can understand you from a distance even if there’s chatter, music, or a movie playing on the television. Echo’s brain and voice belong to Alexa––kind of a souped-up Siri––whom has access to more than 3000 ‘skills’, largely from third-party brands. The skills are rather like apps for Alexa, they teach it things to teach you, as well as connecting with and controlling your smart homes devices. Just say the magic word, ‘Alexa’, and issue your commands. Amazon also sells smaller versions of the Echo, the Tap, which is a portable device, and the Dot, which is a mini speaker you can fill your home with.
Released in late 2016, Google’s big entrant in the market is Google Home, a voice-activated speaker that acts as a virtual assistant while linking with other smart devices to stream your favourite music and shows, control lights, and adjust heating and cooling. Supported devices include the Chromecast digital media player and products from Nest, SmartThings, Philips Hue, Honeywell, and Logitech Harmony. Using the intelligence of Google Assistant, Home is capable of two-way conversations and can offer translations in more than 100 languages. While the Echo has it over Home in the skills department, Home is better at connecting with your television due to its relationship with Chromecast. Just say, ‘Hey Google, play Stranger Things on the living room TV’, and hey presto, you’ll be in Hawkins, Indiana.
There are other players in the virtual assistance market. Apple is forging ahead with its Homekit system which makes use of Siri, although it relies largely on an iPhone or iPad to input controls rather than through a Wi-Fi-enabled microphone/speaker. If you have a Windows 10 device you won’t be talking to Siri or Alexa but Cortana, Microsoft’s own version of a virtual assistant. Cortana links with more than 1000 apps and services, but is being pushed as more of a personal organiser than a smart-home platform. But despite all the improvements in ASR software, there’s still a way to go before we’re having meaningful conversations with our computers. We’ve gotten better at teaching computers to understand words and basic commands, but the next great leaps will need to be in artificial intelligence––the ability to comprehend the context, emotion, subtlety, and nuance of human speech. There are significant obstacles to achieving this level of understanding, but scientists from San Francisco to Beijing are working on artificial neural networks (ANN) and feeding them millions of pieces of data, images, and videos in the hopes that they will learn to think like us.
Ramifications for SEO
The emergence of voice recognition software won’t just change the way we live our day-to-day lives, it will also fundamentally change the way the internet itself works. Think about how you’d normally find a place using Google. You might type something short, like ‘MGM Grand location’, into the search window. But how would you ask for that same information verbally? You’d probably phrase it as a question, such as ‘What is the street address for the MGM Grand?’ It’s a very different way of inputting a request. As a result, Google and other search engines have had to change the way their algorithms work in order to learn and interpret these natural (to us) phrases. As more and more people use voice searches, this has enormous flow-on effect for anyone who uses SEO to influence search results. SEO writers will move away from short key words to long-tail phrases and sentences, particularly ones that begin with words such as who, why, where, what, and how. Failure to respond to these behavioural changes will cost businesses impressions and, ultimately, money, so it’s vital to have a marketing team that understands the challenges presented by the rapid adoption of virtual assistants and other voice-based technologies.