The Voice Assistant War Started at CES - But What Does It Mean For Our Privacy?

By James O Malley on at

Each year at Consumer Electronics Show the challenge for technology journalists is to look at the thousands of products on show and figure out what it all means for the industry, and our world. Each year, there are clearly trends. In previous years, wearables and VR have both walked away as the big trends of the show. So what is it going to be this year?

One fairly early bet appears to be that voice assistants have come of age. In just the last days we've seen Amazon's Alexa assistant appear in non-Amazon devices for the first time, such as LG's smart fridge (yes, that's a thing). Huawei has packed Alexa into one of its latest phones. Google's rival technology, the Google Assistant, has also made its way into Android TV devices, such as the Nvidia SHIELD for the first time. And Ford has even packed Microsoft's assistant - Cortana - into a car.

What's clear is that the industry is betting big on voice technology, meaning that over the next few years we're going to be spending more time talking to our devices. We'll also be embedding many more microphones into the objects we use in our daily lives. The next big battle in the tech industry, no doubt, is who can make their assistant the most accurate the most convenient to use.

But noticing this, I didn't start to think about Google versus Amazon, or Microsoft versus Google or whatever. It made me think about the amount of privacy we're trading for the convenience of voice control. Will there ever come a point when we decide that our listening devices simply listen to too much?

Drip, Drip

There’s never going to be a single moment when we decide that technology is simply too invasive. Today, we’ve already grown used to our devices listening, watching and monitoring us. Simply look back at recent history to think about how much of our previously private lives we're willing to share.

12 years ago, we traded the ability to read our correspondence to Gmail, so that it could show us adverts. But hey, Gmail is lightyears ahead of every other email provider so it’s worth it.

7 years ago, we traded knowledge of our location for the usefulness of our phones knowing where we are. When Google Now pops up and tells us - without asking - the quickest way to get to the next event in our calendars, it feels a bit like magic.

Around 3 years ago, we started to trade our health data - because hey, monitoring all of health analytics is fun when we can compare it to our friends.

And most recently, we’ve started to trade conversations in our home but the ease of being able to control our smart home and play music, simply by saying “Alexa…”, or “Hey Siri…”.

Of course - the reality is that we’re not just sharing this information with our devices, but others, such as a handful of mega-corporations and intelligence agencies. There’s also the ever-present risk that our data could fall into the hands of hackers.

And this leaves me wondering… what’s the next privacy invasion going to be? What will be silently give up next, because in return we get some sort of convenience? My guess is that many of us have already got the devices in our homes that are going to enable it.

Adding Context

Voice assistants are already changing how we live. If you’re an early adopter and have picked up an Amazon Echo, you’ll recognise the utility - the ease by which we can bombard Amazon’s intelligence with questions. What’s the weather like? What’s on my calendar? Turn on my lights. Play me some music. Annoyingly, it's brilliant. And the products we're seeing come out of CES demonstrate how this sort of thing will shortly become ubiquitous.

But at the moment there is a big limitation to the technology: They only know as much as we tell them. Though they are always listening, to trigger them you must use the trigger word ("Hey Alexa!") - and only then will they send our words off to a server farm for processing.

This is perhaps at odds with the ultimate goal of these voice assistants: To make using them as easy as natural conversation. Because Alexa can only hear what you direct at it, it will have no context from which to pick an answer - leading to stilted, still somewhat inhuman conversations. If you turn to another human conversation partner they’ll have been listening along - whereas with voice assistants you have to start all over again.

So surely, the next logical step for voice assistants - and the next drip to fall in the drip-drip erosion of our privacy - is for them to start listening to everything?

Amazon, Google or whichever company comes up with it first won’t phrase it like that, of course. They’ll think of a brand name… VoxContext™, or something like that - and give it a friendly logo. And it will, annoyingly, be more useful than what existing technology offers. Imagine half way through a conversation simply being able to say “oh yeah, buy that for me, Alexa”, or “I’ll have to Google that”, and have the assistant know exactly what you’re talking about from the broader conversation.

"It looks like you're trying to argue with your spouse, would you like any help with that?"

The tech companies have already introduced some context sensitivity to their voice assistants, at least for the action taking place on screen. Google has already demonstrated an understanding of context with its Assistant. For example, if using its Allo chat app, mid-conversation you can call on the assistant to, say, book a table at a restaurant.

Siri can already understand context too. If you’re looking at something in Safari, you can say to Siri “remind me about this when I get home”, and it’ll flag up the same web page when you reach your house.

Given these devices are already reading what’s on our screens - isn’t listening to us chat the next logical step? Wouldn't it be useful to say "Alexa, add that to my shopping list" mid-conversation with your partner, without having to specify the specific product that you're referring to, because Alexa will have been listening all along?

Your reaction to the idea of this might well be horror: It’ll mean that we would willingly sharing literally everything we say with the cloud - and listeners unknown - in order for it to work. It’s easy to write terrifying, dystopian fiction about how such a technology could be successfully used by an authoritarian leader: Just imagine how much power someone would have if they could listen in, in full, at any time to whoever they choose.

But the weird thing is that however creepy it might feel now - if CES points to the next battle being over voice control, this feels like an inevitability.