Did it ever occur to you that since not that long ago, at any given moment, we are being constantly listened to by dozens and hundreds of electronic ears. I am primarily talking not about our very own personal buddies Alexa and Siri, but rather about a diversity of less identifiable smart devices and apps of our friends, co-workers, fellow commuters, and occasional passers by. Talking about a trip to Spain over a lunch with your friend, only to find numerous ads of Andalucian seaside villas in your news feed the very next day, became a new norm and doesn’t surprise us any more. But talking about the same trip with your phones off and still getting those villas pop up in your search results – is it something we can expect in the near future?
Voice and speech recognition systems are only getting better and cleverer, involving cutting-edge AI and machine learning technologies. The apps that listen to us and the massive data centres behind them are not that primitive pattern-based monkeys anymore. They are learning to understand more languages, more intonations; they are learning to identify the mood and the reasoning of the speaker; they are learning to decode Yorkshire dialect and Cockney slang. And given they have enough storage capacity, they can totally hold voice patterns for every person they are aware of – and use that database to identify unknown speakers on any recordings they get their hands on.
What this ultimately means for you is that a random okaygoogle on your fellow tube traveller’s phone can overhear you chatting with your girlfriend, identify both of you by the patterns of your voices, recognise what you were talking about, and use that information to spam you.
This indicates a major shift in paradigm. For many years, we used to be in control of most information flows between us, people, and computer systems. We encrypted sensitive information. We restricted access to critical computer systems. We assigned labels to data to separate sensitive stuff from unclassified, and controlled the transition from one classification level to another. We could control all the milestones of the data we possessed, from its genesis through its useful lifetime and to its disposal. We could apply verifiable security measures to each such milestone and be sure that our data remained adequately protected at all times.
Today, we are not in that control anymore. Our information is slipping away from our hands, being captured all the time without us even knowing about it. You can assign the highest sensitivity level to your strategic roadmaps and encrypt them in transit and at rest, but you will never know if your coffee maker made a note of your private telephone call with your CFO about them this morning. You can turn Siri off and even throw away your phone, but you won’t have the slightest idea of what sorts of smart TVs, fitness trackers, or Furby Booms are going to be quietly recording your verbal interactions and when. Walls are naturally getting their ears.
And this means that we should be more careful about what we are saying out loud, not only in private but in public places as well – especially in public places. Anything you say may be used against you. Old school methods, like going out to the fields to discuss a confidential matter tête-à-tête, might well start looking for a new life. Sign languages will gain huge popularity in near future, too.
Not sure about you, but I’m definitely signing up for the course.