I have had two occasions where it has become clear that my Google Home is listening and acting on things that are not said to it directly:
1. Much to my confusion, I received a push notification about the weather in Ibiza. It only became clear why a week later when my wife told me she was going there with her Mum. She had been discussing it next to one of our two Google Home devices.
2. Just now, I had a meeting where the word "Glasgow" was said several times (again, next to my Google Home), along with discussing an upcoming meeting there. When I opened Google Maps after that, it opened centred on central Glasgow. I live at the other end of the UK, there's no other reason for it to do that.
Now, I understand that both of these things were "helpful" as Google's algorithms saw it. I appreciate that it's trying to help me. The problem I've got is that as far as I know Google swears it doesn't listen when not addressed directly, when it clearly absolutely does. If it does for purposes that are "good" for me, what else does it do with what it hears.
#google #googlehome #googlenest #voiceassistant #dataprivacy
@VoxDei right - but it is ALWAYS listening for the wakeword, and lots of things have a similar enough waveform to the wakeword, that it sometimes get activated. For example do your best baby voice impression to it "ohay goo goo" it will still respond.
As a result, it will get triggered many times when it was not intended to be triggered.
There is an episode of #MurderSheWrote that will trigger mine every single time, there are some gangsters in a pool hall and even though they don't
@VoxDei say anything even remotely close to anything that sounds like the word Google - it will still activate every single time. So either it's the background music or something else - but clearly it "hears" a waveform that matches close enough to what it is listening for.
@VoxDei Don't forget that when you are thinking of travelling, somebody is putting in a lot of search requests. Just say Moscow to it ten times...