Love it or hate it, your voice contains a ton of information about you. With voice assistants getting smarter day by day, we are looking at a future where we would be easy prey for manipulative marketing.
Can you tell how your friend is feeling just by hearing their voice in a phone call?
Most likely. It's hard to articulate how we do that, it's just one of those things.
The reality is that our voice contains information well beyond the content of our speech. Unless you’re listening to a certain politician, then the content does not contain much information either. But let's not go there...
So. With the emergence of voice assistants, we are getting closer to a world where the Siris, the Alexas and Google Assistants will all be able to tell if you are upset, sick, depressed or about to get a divorce. What? But how ready are we really for a future where “Her” is not just a movie plot, but a real summer wedding you attend?
Voice Analysis: What We Know Now
With every new study, we are decoding more information in the human voice. Researchers now know that when we speak, our voice includes information about our mental, and physical health and even the health of our romantic relationship(s).
Changes in the tempo of your voice could be significant markers of depression. Other qualities in your voice undetectable to human ears could give doctors clues about your coronary health or the likelihood of you getting Parkinson's disease.
And – this is a little frightening – 'information' in your voice can even tell you if your relationship will last. In a study published last year, computers analysed the voices of couples in therapy over two years without understanding the content of their speech. But, they were able to predict with 79.6% accuracy which couples would end up divorced! A bit worrisome given that those we rely on today – human psychologists – were correct in only 75.6% of the cases.
Turns out, what was said between the lines was more telling. Just imagine how awkward it will be if Alexa recommends you the top-ranked divorce attorney in your city when all you asked was: “Top-ranked Mexican restaurants nearby.” Something is getting unplugged for good.
Voice Tech: Where We Are Today
Before we dive deep into the pool of possibilities, let’s take a look at the current state of the voice assistants.
The market for them is growing rapidly and they’re readily available on our smartphones, for our homes, in our offices, and cars. It is not a stretch to say that they will be an integral part of our lives in the future. For many people they already are. However, they are still a few, slight adjustments away from being totally useful.
For instance, voice assistants are still quite bad at understanding context. They have no clue why the user asks a certain question and how it relates to their previous inquiries.
Another challenge is that our world is still not as connected as tech companies would like it to be, making it less useful for users to control their tools and environment.
Finally, all the major voice assistants belong to Silicon Valley giants: Google, Amazon, and Apple. Google, with its multilingual expansion, and Amazon, with its shopping focus, are the major players in this arena. So, they're interested in greater connectivity. But why? To better connect us is not an answer, it's a tautology. But that's what most of their arguments seem to amount to. We suspect money – theirs not ours – might have something to do with it...
Voice Tech: Where We May Be Headed
Regardless. If the roadblocks mentioned above can be overcome, an interesting future awaits us. An all connected world, where your voice assistant has real time access to all of your online activity and can tell how you feel only from the tone of your voice might sound creepy. And it is.
Unfortunately, for those of you who are freaked out, Alexa, and Google Assistant not only store your voice history, but also the actual recording of your voice. The reason, technically, is so that when the system doesn’t understand something, you can retrain it.
Here is the thing though, with all the information stored about you, the voice assistant can be a highly effective, highly precise marketing operative. It will know where and when you will need something and alert you. They already send notifications. The assistant can even change the “emotional” tone used to respond to you based on your mood.
True hyper-personalised marketing
Depending on the way you view hyper-personalisation of marketing, you may find the future promising or, very dark. Has everyone seen Minority Report by now? Furkan (my co-author of this article) is the pro-personalisation guy, and would rather be presented with ads that would catch his interest. But Didem (my co-author of this article) is more sceptical as advertising is fed by research in behavioral economics and psychology.
The advertising industry argues that it's meeting needs. Its critics argue that it is creating desires.
Hyper-personalisation is a whole new ballgame all together. With hyper-personalisation, your mood and search history can be combined to target you with products. Imagine this scenario: you have have a need (desire?) you're not even aware of to impress people (but Google Home knows of course). So it presents you with the perfect opportunity to do just that with a brand new, expensive TV when all your friends are with you. So you, yes you, go ahead and buy it.
One could argue that it will just make life easier – devices bringing you what you are already interested in buying. But perhaps you're just being manipulated into buying things that you don’t really need, worsening your impulsive buying habits given the barriers to purchase is removed.
Remember how credit cards changed our buying habits, introducing a transaction where no real paper money is ever handled. With these powerful tools living in our homes, who can guarantee they won’t prey upon us during our most vulnerable moments like right after getting fired or losing a friend?
Is it all bad?
There may be some benefits to these technologies of course. If we know medical emergencies before they happen, like fatal heart attacks, we can get people the help that they need. If these devices help as many people as possible to get early diagnosis on things like Parkinson disease, we could improve their treatment or quality of life.
If someone is feeling suicidal, the human sounding voice assistants could talk to them and prevent it from happening long enough so there is help coming for this person. Though these positive possibilities are just as realisable, our greedy and violent collective history as humanity points could lead us into another direction. Medical help will probably not be available universally. It would likely still be distributed to the privileged few across racial, economic and political lines. As for helping the vulnerable among us, without human connection and knowing someone real is talking, would those in need even be comforted in the least?
What about the Money?
As with most things, after the scientific breakthrough comes the commercial application. Companies like Sonde Health, Peak profiling and even IBM’s Watson is in on the game. These companies are trying to commercialise the research around voice as a biomarker of disease and promise early diagnostic results.
There are already startups that promise to deliver vocal pregnancy tests and MIT Researcher Rebecca Kleinberger says at this rate, it is not a big stretch to imagine computers will know we are pregnant before we do.
Some efforts like Cogito are being backed by the US Department of Veteran Affairs while the US army is partnering with MIT to develop tech that can detect brain injury from a soldier’s voice.
Duplex showed us how human-like the assistant can sound, but the question of whether it is something we want/need or something we should think seriously about putting on mute. Technology and the companies building it seem to be making all the decisions. Perhaps it's time to start talking back. Maybe then, our voice assistant might just pause to listen.