in

‘Hey Google, stop eavesdropping.’ Is the free smart speaker an intrusion on privacy?

When Spotify announced it was giving away a free Google Nest Mini smart speaker to all new and existing paying subscribers at the beginning of the month, it sparked a frenzy.

While the streaming service had planned to run the offer throughout September, all stock had been claimed just four days later as users rushed to snap up the offer. Although Spotify didn’t respond when asked how many of the £49.99 speakers they’d shifted, it’s safe to say the giveaway was a success.

But, as everyone knows, there’s no such thing as a free lunch (or smart speaker). The devices have become increasingly controversial in recent years, following reports of humans listening into snippets of recorded conversations, their data collection methods and even accusations of spying on users.

The Spotify/Google deal triggered an avalanche of jokes on Twitter as users quipped about turning their homes into a Black Mirror dystopia and how the “government can hear me more clearly now”. “Imagine caring so little about your data privacy you jump on getting one of those free Google Nests from Spotify,” added one disgruntled tweeter, before screenshotting an order confirmation and adding: “Anyway I can’t wait to make mine tell me what else that actor has been in every 10 minutes”.

Although it’s impossible to know whether the impetus behind the giveaway was Google was looking to shift a load of unsold stock or a blatant stunt to get more Nest Minis in our homes (or a combination of the two), the stunt is an interesting illustration of our conflicting attitudes towards smart speakers.

Amazon was the first to lead the charge with the launch of the Amazon Echo in the UK in 2016, much to the chagrin of any girl named Alexa, and was swiftly followed by Google’s Home and Nest Mini (formerly Home Mini) and Apple’s HomePod.

All three work in much the same way – an AI digital assistant answers questions with information drawn from the web, playing music on request and controlling connected devices, including smart lamps, TVs and kettles – and have spawned a plethora of cheaper knock-offs around the globe.

Soaring popularity amid privacy concerns

While Amazon, Google and Apple always resist officially confirming how many devices they’ve sold, analysts are confident that global smart speakers sales had crossed the 200m mark by the end of last year and estimate the market will be worth a whopping $35.5bn (£27.2bn) by 2025. One in every five homes in the UK is estimated to use one, and they have been praised by disability campaigners for enabling less able-bodied people to interact with technology in simpler ways, and for helping to combat loneliness in older users.

All major smart speaker makers have been rapped for not making it clearer they hired human workers to transcribe user conversations (Photo by Jeenah Moon/Bloomberg via Getty)

The tech giants claim that being able to speak commands aloud to a device “removes friction” (Silicon Valley parlance for ‘makes things easier’) by eliminating the need to press any buttons or type ourselves, but voice control also ushers in a new way of interacting with technology that encourages us to view products as sentient beings.

The Centre for Data Ethics and Innovation published a paper last year suggesting that the practice could change the relationship vulnerable people, particularly children, have with technology, adding that “the risk is that such relationships could be exploited resulting in people passively accepting the responses and instructions they receive without considering their own safety and security.”

It also highlighted how the speakers’ recordings could provide the manufacturers with “new troves of data” to create profiles of customers for advertising purposes (something Apple has fiercely denied) that could prove an attractive target for hackers.

This brave new world created by digital assistants and the speakers they’re housed in has cleaved us into three camps: the evangelists, the nay-sayers and those who aren’t entirely on board with their data collection but don’t care enough to not use them (or deny themselves a freebie).

A close up of Amazon’s Echo Studio smart speaker (Photo by Neil Godwin/Future Publishing via Getty)

With opinion divided within the tech security industry, it’s little surprise the general public is unsure about how to feel about them. While the tech giants are at pains to hammer home how Alexa, Siri and Google Assistants wait for their ‘wake’ words to start recording instead of constantly monitoring what you say, the revelation all three companies were paying human contractors to transcribe snippets of conversations to improve how the AI understands human speech last year did little to consolidate trust in either their products or their intentions with our data and triggered a wave of new privacy protocols.

Users have also reported many instances when the assistants start speaking or even laughing unprompted, suggesting they’ve been triggered by something other than their actual wake word.

‘Just not right for companies to collect data this way’

“I absolutely refuse to have one in my home,” says Aimee Davis, a social media manager. “You’re actively inviting a listening device into your home, why would anyone want that?”

Johnny Bethea, a marine geologist, agrees. “I just don’t like the idea that there is the designed capability for them to collect data passively, without your knowledge, he says. “It’s just not right for companies to collect data this way; if it was your neighbour looking into your spaces through a peephole, would that be okay?”

Facebook’s Portal smart speaker was poorly received in the wake of the Cambridge Analytica data scandal (Photo by Facebook)

Ultimately, how comfortable you are sharing a home with a smart speaker is likely to be down to whether you feel its benefits outweigh the potential downsides. Catherine Cooper, a freelance journalist, says she loves her Echo and is unconcerned by what it may or may not overhear. “To be honest, I don’t really care if Amazon know I’ve run out of potatoes (we mainly use it to compile our shopping list),” she says.

Similarly, Hannah Thompson, a meditation teacher and success coach, has an entire home full of connected devices she controls by speaking to Amazon’s Alexa, including turning on the TV and starting up Netflix. “It’s very handy if you’re cooking and carrying plates out of the kitchen and you want to turn the lights off as you go, but you don’t have a free hand,” she explains. “Although she doesn’t like my voice compared to my partner’s, so I often have to repeat myself, which feels mildly annoying.”

Whether the host of Nest Homes wending their way to their new owners end up becoming useful tools or gathering dust in drawers, speaking to our devices is only going to become more and more normalised. The most effective solution to combatting smart speaker data collection is to either mute its microphone while you’re not using it (a simple procedure on all three major models) or to not use one at all.

How smart speakers work

Smart speakers all work in roughly the same way. “At any given moment, the device holds a continuous ‘buffer’ of the last few seconds of sound recorded from its environment, which it scans for the wake-word,” explains the Centre for Data Ethics and Innovation.

The wake word varies from product to product, ‘Alexa’ for Amazon’s Alexa, ‘Hey Siri’ for Apple’s Homepod and ‘Okay/hey Google’ for Google Home/Nest Mini. Once the wake-word is detected, the device begins recording and streaming audio to the cloud for analysis and storage. In the cloud, the snippet of audio is converted using speech-to-text technology, translating it into a structure that machines can derive meaning from.

This information is then passed to a dialogue manager that selects the most appropriate response to the request (e.g. playing a particular song, relaying the weather forecast or searching the internet for the answer to a question). This may need contextual information, such as the device’s location when considering weather data.

The system then readies the spoken response to the request (“Sure, here’s Paint it Black by the Rolling Stones” etc) and streams it back to the device, alongside any other actions or additional questions to complete the request.

Brian Brackenborough, a chief information security officer at a major UK broadcaster, owns around 10 Echo devices and feels the experience can be likened to using social media – it’s an informed risk.

“For me, it’s about making informed decisions. I know what ‘could’ happen in the worst-case scenario, and as such, weigh up the pros and cons,” he explains. “My Alexas provide me with more benefit than any risk that I can think of.

“If you accept that what you share with your smart speaker ‘could’ end up more public than you anticipated, but then make the decision to do it anyway, then that’s on you.”



Source link

Diary Of An Apprentice Astronaut review: An enthralling book

Raiders reportedly under investigation by NFL for allowing unauthorized locker room access