The search giant wants to beat Apple and Amazon to a voice-enabled interface.
Even though it was billed as a hardware event, Google’s product launch Tuesday was really about artificial intelligence, which is now under the hood of almost all the new gear coming out this year.
The power of Google’s computer brain is a result of all the searching you and everyone else in the world have done on Google over the years. And now the company is using all that data to build a voice-enabled personal assistant, which will debut on Pixel, Google’s new smartphone, as well as Google Home, the company’s answer to Amazon’s Echo.
Google CEO Sundar Pichai said they’ve gotten really good at helping people figure out what they want to know. Google’s Knowledge Graph, that small box on the top of search results with the quick answer you’re looking for, now holds 70 billion facts, all of which help power its new AI assistant.
Google isn’t the only company vying for your spoken commands. All the big players, from Apple to Amazon to Facebook, are investing in the idea that voice is going to be the next digital interface. And they’re all racing to become the de facto standard for voice-enabled AI. Google won desktop search. Apple won mobile with the iPhone, and Facebook won social media. But voice is still up for grabs, even if Amazon is off to a head start with the popularity of Alexa, the voice-powered AI that lives inside the Echo.
Google’s image recognition abilities are rapidly improving in accuracy, too. Sundar says Google’s AI can now identify images with 93.9 percent accuracy, up from the 89.6 percent accuracy they achieved in 2014. Sure, it’s fractional, but it amounts to dramatically refined results.
And when it comes to natural language processing, Google says that its AI is approaching human levels of understanding. Now that Google Translate uses machine learning, their accuracy clocks in (on a scale of 1-6) at 4.263, said Sundar. Humans generally score 4.263 on the same scale.
Here are all the ways Google is using artificial intelligence to power its new hardware offerings.
The company’s Pixel Phones will be the first devices with Google Assistant preloaded. It’s either voice activated or turned on with a long hold of the home button. The idea with the assistant is that it glides from search queries to text to your calendar, all with voice command. Yes, it’s similar to Siri, but it’s fair to say Google is better at search than Apple, so its assistant should reflect that.
The Pixel also uses AI to power an intelligent burst mode in the camera, Smartburst, that takes multiple shots and chooses the best.
This is probably the most important product the company shared today, as it highlights the company’s vision for the future of computing. The company packed its artificial intelligence into a tabletop speaker that resembles an air freshener.
Google Home is voice activated, can handle complex queries and is programed to learn your preferences. So the more you use it, the smarter it gets.
The new router system creates a mesh network throughout your house and uses artificial intelligence to take some of the guesswork out of picking the hotspot with the strongest signal.
Even the video streaming stick is benefiting from the company’s AI push. Google showed how Chromecast can work as a smart endpoint in conjunction with Google Home, allowing a Chromecast-ready television to broadcast the results of queries like, “Show photos from October” or “Play John Oliver clips.” If it doesn’t get it right the first time, correct it and next time it should know better.
The one outlier of the bunch was Daydream, which is part of a separate push to make Android the operating system for virtual reality. We’ll have more on that effort in a bit.