Smart home hubs are continuing to evolve, and Google just added a pretty important feature to its own hub, the Google Home.
Previously, Home only linked up to the account of whomever set it up first. Now, the device will be able to handle multiple accounts and tell who’s speaking to it, offering personalized answers to some questions.
That’s a feature that Amazon’s Echo doesn’t have. And it’s important for a voice assistant that’s designed to run your household. For an assistant such as Siri, which lives on devices used by just one person, multi-account support isn’t as important. But home hubs sit in a central location and operate things such as your lights or your thermostat that everyone will want to be able to control.
Being able to identify an individual’s voice may also help cut down on some unwanted surprises. Google said in a statement that the new feature makes it so that “only you would be able to shop via Google Home.” So others — i.e, your kids or an intelligent parrot — shouldn’t be able to tell Home to buy something on your account. That avoids instances like one in San Diego this January when Amazon Echo units started ordering dollhouses after hearing a news anchor said “Alexa ordered me a dollhouse.” The anchor was reporting on a story about — what else? — a child buying something without permission on the Echo.
(Amazon chief executive Jeffrey P. Bezos owns The Washington Post.)
That said, the company also noted the voice feature isn’t yet foolproof. “We’re just getting started and we won’t be perfect,” the statement said. “We don’t recommend that users rely upon voice identification as a security feature.”
While there are benefits to having home hubs distinguish different voices, there are also some privacy implications to think about. Consumers worried about their voice data being collected in general should think twice before picking up a home hub, said Bradley Shear, of the privacy-focused, Maryland-based Shear Law. With this new feature, he said, it’s worth keeping in mind that Google will have even more specific information about you, which could be used in ways that consumers may not realize — particularly if it’s combined with other information tied to your Google account.
Consumers should also think about how this information could be used outside of the company, Shear said. He pointed to a recent murder case in Arkansas where police asked Amazon for audio recorded from the suspect’s home hub, the Echo. (Amazon originally fought the order before the suspect agreed to share the information.) Shear said that case illustrated how recordings made in your home may end up being used in unexpected ways.
“It has a clean voiceprint from you,” Shear said, speaking of the new feature. “Once something is digitized, you don’t know where it could end up.”
From a usability standpoint, there are also some problems with these hubs that identifying individuals by voice can’t fix. This won’t solve the “Burger King” problem of commercials being able to trigger the Google Home through your television. Anyone in range of the Google Home will still be able to use its non-personalized features. So even if you don’t create an account for your guests or your children, they’ll still be able to ask it to do things such as answer questions, set timers, or play videos on a Chromecast.
To set up the new feature, users will have to hop in to the Google Home app, which should have a new option for “multi-user” support on any connected Google Home connected to your network. Home owners can add up to six accounts, according to a company blog post.
Google Home owners will have to teach the hub how to identify their voices by saying “OK Google” and “Hey Google” twice into the device, when prompted. This allows it to learn different characteristics of each person’s voice. This kind of training should be familiar to anyone who uses Siri, or the voice assistant on Google’s phones — the key difference being that you’ll have to go through it multiple times on Home, depending on how many accounts you want to hook up.
Google said that it will use its network to “compare the sound of your voice to its previous analysis so we can understand if it’s you speaking or not.” The company promises this should only take a few milliseconds.
The feature is rolling out to all U.S. users starting Thursday. Google said it also plans to bring this feature to the United Kingdom in “the coming months.”