Well you have to train the system some how when it’s expected to work across a wide range of accents and speech patterns. At some point people need to recognize that these devices are always listen (that’s the only way they can respond to their name being called) and sending that info back to the cloud.
I wouldn’t be surprise if Siri and Cortana do the same thing.
I’ll never buy one and I don’t use Siri on my iPhone cause I figure that is what has to happen to make them work as well as they do.
