I honestly think the way this works is a lot weirder than what we might imagine. We’ve had pretty strong machine learning for a while… we still don’t know exactly how it works. The likes of Google ingest all manner of disparate data points, buying and selling them to each other, responding to the push and pull of markets and trends. We feed these entities continuously with our actions, whether we know it or not, and it gets digested in ways they might not even know about, picking up patterns that humans would never think to look for, and then feeds it back to us.
These players have a realtime profile on every single person who touches their platform, and users continuously touch that platform with almost everything they do. They can map all the intersections of those profiles, the rates of change
they don’t need our microphones.
Why would they need to listen? That’s a lot of processing power
But… why would they care? It’s your energy bill, not theirs.
It definitely totally happens every single day but nobody ever manages to get a piece of hard evidence.
Are you just lazy and in denial? There is traffic sent from both Alexa and Google Home every time it detects something that it considers speech, regardless of whether the wake word was detected. This is not debatable, set up your own tests.
I have Google Home and monitors on my own network, I’ve never seen suspicious activity. Here’s somebody else doing it: