There’s a kind of revolution underway in voice-activated AI, according to Expect Labs’ Tim Tuttle in an article for TechCrunch. Tuttle says that “voice recognition in machines is getting very good and is going to get so good that it will completely change the way humans interact with their computing devices.”
Tuttle argues that over the past year and a half, speech recognition technologies have gotten 30 percent better, an improvement that he says exceeds the sum total of the past 15 years’ technological advances. Tuttle attributes this to deep learning systems such as DeepMind and Vicarious, and argues that as more users interact with these systems and more data thereby becomes available, the systems’ accuracy will constantly improve. He also argues that more companies are working on this kind of technology in a variety of applications, especially in the Internet of Things. That is undeniably true; we’re already seeing the emergence of systems like Dragon Drive, a smart car-based AI system designed to use voice recognition to personalize the driver’s experience.
Tuttle goes on to argue that soon AI systems will be listening to users constantly, asserting that a connected “conference room, automobile or wearable device that can listen to our conversations and understand what we need will eventually become the norm.” That too seems pretty on the mark, as we’re already dipping our toes into such a world with the emergence of such devices as Samsung’s newest smart TV, whose ability to constantly listen in on users is already causing quite a fuss among privacy advocates.
From there, Tuttle goes into some more speculative territory about the possibilities of artificial intelligence and an eventual “Robopocalypse”. Whatever the future holds in those areas, he is seems to be right about voice recognition technology’s more immediate future. And as more advanced technology allows for ever-smaller deployments, it might soon be hard to get away from the various everyday devices that are listening to us all the time.