The age of sentient technology
Amid the excitement over Amazon’s latest product launches, which I anticipated in a previous post, one aspect of potentially radical importance has gone largely unnoticed: the idea of equipping these devices an intelligence that goes beyond simply understanding words or phrases, and instead being able to understand nuances such as the tone of voice of the user, interference or other things they might hear that could indicate certain situations.
For example, by whispering something to our assistant, the device would understand it was required to reply quietly, so as not to disturb other people nearby. Perhaps home assistants could be taught to distinguish if we’re angry or calm from the tone of our voices. Perhaps they could activate certain functions if they detect anything that may indicate an emergency was taking place in our homes.
What would it mean to equip home assistants with a layer of intelligence that made them “sensitive” to what’s going on around them? Allowing our assistants to understand what we want by screening our requests from background noise such as a television or conversation is a step forward and would certainly increase customer satisfaction (make the mistake of placing your Amazon Echo or your Google Home near the television and you’ll know what I mean). But what if we add another layer: a device that were able to distinguish the sound of glass breaking glass or detect an unusual increase in temperature or humidity and then automatically activated a camera, sent a message or notified the police or firefighters? The possibilities are endless, and so is the potential for things to go very wrong.
Actually, an algorithm is perfectly capable of recognizing the sound of a broken glass. The challenge is for it then to listen and work out if somebody has simply dropped a bottle or if it is a thief who has entered through a window, which requires very sophisticated intelligence. Such services could be considered a convenience, providing additional intelligence that might be useful in certain cases and that most of us would be happy about having. But that layer of intelligence could be used for a range of security issues: women threatened by their partner, for example, could introduce an algorithm that detects
certain kinds of language, tone of voice, etc and discreetly alerts the police. What are the possibilities involved here?
The third layer is a little more complex, and yet in all likelihood, within a short time, it will be part of my day to day life: a machine that understands my state of mind, my level of stress or my immediate circumstances when I ask for something. In other words, if I’m frustrated, in a hurry, if I’m being ironic or if I’m looking for a serious answer. I find it rather irritating when Apple’s Siri believes she has detected irony and answers “good question”, but then I have not yet acquired the habit of “talking” with my robotic assistant. Will a time come when I do? I confess to having dedicated time during traffic jams to asking Siri for things and trying work out her answers, what logic they are based on. Home assistants have been programmed to respond to jokes and seemingly to understand what we’re saying, but obviously, these devices are still not capable of holding a conversation. At the same time, as we know, we have adapted to technology in different ways: many of us keep the television or radio when at home, even if we’re not listening, just to keep us company. What might we explore in this area?
How about a conversation with our home assistant to discuss a particular subject, during which it detects, by our vocal inflections our mood or circumstances? Or being read to, but with the right intonations, rather than a monotone? Another possibility is a resume of the news whereby we could ask the device to expand on a particular topic or seek additional information from other sources? Home assistants could even draw our attention to particular news items it knows are of interest to us.
If we put home assistants capable of listening to us at home, in the car and at work, the question of making the most of their possibilities comes down to a fear of loss of privacy. At the rate things are going, it is only a question of time, of adoption dynamics and the creation of new applications before our homes will be filled with these types of devices. The time has come to start taking home assistants seriously.
(En español, aquí)