‘Voice assistants’ have become less vocal, but much more powerful

2022-05-13 20:00:00

“Sometimes Alexa acts on your behalf but you don’t even have to ask, for example triggering an automated action or routine. Sometimes a sensor can recognise you and then surface information that’s relevant. And sometimes you’ll be able to act on your needs by controlling smart home devices with your words or with touch,” he said.

“All of these share the trait that they come together, under the Alexa auspices, and form parts of ambient computing.”

At Thursday’s Google I/O keynote, ambient computing was also a main topic of conversation. Google said it wants to use its device’s existing ability to tell users apart, along with new privacy and security features, to make interacting with the Assistant more natural.

An incoming update for its Nest Hub Max will let the smart display listen to you if you look straight at it and make a request. Google will also give users more control over so-called “quick phrases”, which are common requests the Assistant will constantly listen for like “set a timer”.

While the new tools negate the need to say “Hey Google”, that doesn’t mean the Assistant as we know it is going away. Having a wake word still serves an important purpose, said Google Assistant product manager Jaclyn Konzelmann.


“Sometimes people call my name before they want to talk to me. But other times they don’t. They might simply look towards me. They might tap me on the shoulder, or if it’s just the two of us sitting in a room, they might simply ask their question,” she said, adding that the Assistant should react naturally to these different prompts as well.

And despite the prevalence of automation and home touchscreens, Konzelmann doesn’t think talking to the Assistant would become a thing of the past.

“Speech will always be a very powerful way to get things done. The spoken language is something that almost all people are able to lean into to issue commands,” she said.

“That said, I think that our goal will always be to meet users where they are, and if that means expanding the types of ways that we want to enable interactions, we’re not going to shy away from that.”

Both companies are awake to the concerns about privacy, which are omnipresent in the discussion of smart homes and voice assistants, and especially since many of their products now have both microphones and cameras. Both say they’re committed to processing as much data as possible on the devices themselves without uploading it, anonymising the information that must be sent to the cloud, and including easy ways to block the mics and cameras or delete personal data.

Related to these concerns, Wi-Fi speaker pioneer Sonos announced on Thursday it had created its own voice system, specifically designed to control music and never send recordings or transcripts to anyone. Many of the company’s speakers have microphones, and users have long been able to choose between Google Assistant or Amazon Alexa. But Sonos product manager Scott Fink said the company wanted to add its own third option that had no privacy trade-offs.

Sonos Voice Control can initiate music from most services, as well as control any music by playing, pausing or moving to different speakers.

Sonos Voice Control can initiate music from most services, as well as control any music by playing, pausing or moving to different speakers.

“Some of our customers have expressed concerns about voice privacy, they’ve been dissatisfied with the accuracy, speed and ease of use of existing voice services … and therefore have been happier to choose a product without microphones,” he said.

“That’s one of the things we’re hoping to transform here, where we’re focusing on speed and reliability and accuracy, and something that’s entirely private.”


The voice control can do just about anything you generally use the Sonos app for, although at launch it will not be able to initiate music from Spotify. Summoned by saying “Hey Sonos”, it’ voiced by actor Giancarlo Esposito and will arrive as an option for all voice-enable Sonos speakers in the US on June 1, but the company is yet to announce Australian plans.

Fink said the system works conversationally and didn’t require precise syntax. You can ask for “a little more” after asking for a volume increase without repeating the wake phrase, or you can ask for a vague vibe rather than a specific artist or song, which would generally result in one of Sonos’ many radio station playlists.

But he ruled out the ability to ask for help beyond music-playing.

“It’s not a general purpose assistant, and it doesn’t aim to be,” Fink said.

“It does what it says on the tin, and nothing more.”

Get news and reviews on technology, gadgets and gaming in our Technology newsletter every Friday. Sign up here.

#Voice #assistants #vocal #powerful

Source by [earlynews24.com]