It's not, but the burden of proof is still there. It's up to the claimant to prove that some activity exists, not for it to be assumed to be true and for others to prove it's false.
>But more importantly: at some point the quantity of individual anecdotes, along with the technical ability and interest of the parties involved, become data in and of itself.
Ever heard of the saying "the plural of anecdote is not data"? I think this applies here. Specifically, I think this sort of phenomena is highly susceptible to publication bias/p-hacking. ie. you only remember/hear about all the instances that confirms this theory, but not all the instances where nothing happened. With billions of facebook/google users worldwide, and dozens of possible "topics" being generated per day, I'd expect millions of anecdotes to surface just by random luck alone.
Finally, if we're going by "a lot of anecdotes is proof in and of itself", what does this say about other paranormal phenomena? eg. ghosts or bigfoot? Those have a lot of anecdotes as well.
> It's not, but the burden of proof is still there. It's up
> to the claimant to prove that some activity exists, not for
> it to be assumed to be true and for others to prove it's false.
Isn't that the difference between gravity and "facebook is surreptitiously listening to our conversations", that the former has loads of high quality evidence backing it up and can be trivially demonstrated and the latter doesn't? If someone claimed there was an invisible force that we don't know about, but you can't demonstrate it and the only evidence we have of it is various anecdotes, I'd be skeptical as well.
>Leaking a single keyword per day or per week would greatly enhance such systems' capabilities yet remain very very difficult to detect. Perhaps 52 new keywords per year is enough for them.
I suspect you're getting some of the threads mixed up. Other commenters have mentioned about this sort of activity being detectable via network monitoring, but I never made such claim. Instead, I claimed that such activity would be detectable by the operating system itself, due to the microphone activity indicators. That's not something you can sneak one word at a time, because you need to be listening all the time to pick up keywords.
Her phone is a Samsung A50 probably Android 10 or 11.
> we only have years of anecdotes but zero such trials.
There's some quip about lack of evidence not equating evidence of lacking. But more importantly: at some point the quantity of individual anecdotes, along with the technical ability and interest of the parties involved, become data in and of itself.