Before the Plague hit it was a big discussion. The claim was that if you talked about something (like shrimp or plate o’shrimp) Alexa and all the rest would overhear you and send you ads. Then 2020 hit and I sort of forgot about the whole thing.
Was that ever proven, or what?
How very odd. This was a Big Deal a year or two ago. It was all over this board and the Internet generally. But we have had so much else hit us of late, this sort of went away. Further, a quick search shows no definitive conclusion.
Repo Man?
No explanation. No point in lookin’ for one, either. It’s all part of a cosmic unconsciousness.
Very intense.
The short answer (because there are caveats) is that no major smart speaker should be sending data about what you’re talking about without first being actively engaged by the user with a wake phrase. Sometimes a device will activate if it hears something it mistakenly recognizes as the wake phrase or there is a bug in a software update, but the point here is that while all the devices are constantly listening to you all the time, they shouldn’t be constantly recording or sending any data about you.
Once you’ve engaged the device, however, and depending on what your privacy settings are, Google and Amazon, for example, will collate that data with your individual account and there is a possibility you will see personalized ads that reflect what you’ve said to them.
Apple, on the other hand, assigns your Siri data a random identifier and doesn’t associate any data you send with your personal iCloud account. Apple doesn’t build personal profiles on you for advertising purposes to the extent Amazon and Google do, so even if they wanted to send ads to you, the random identifier pretty much makes that impossible.
So there was nothing to this idea? It was widely believed.
If you’re only talking about when these devices are listening to you, then yes, there is nothing to that idea. Barring any bugs, glitches, etc, no reputable device is relaying any data about your private conversations to anyone without your go ahead. If you’re sitting in your living room next to your Alexa and someone mentions Pussy Whip, and then later you see an ad for Pussy Whip, it was not because Alexa was eavesdropping on you talking about the first dessert topping for cats.
Which is not to say there haven’t been genuine privacy concerns with these devices, some more genuine than others. I get the feeling that might be more of what you’re remembering.
I recall there have been a brief scandal or two when it was discovered that one or another of the major voice control gizmos had been triggering far more than expected or advertised and whatever was recorded and sent to the central servers had been kept and was at least “over-used” if not exactly “mis-used”.
Those early events have not recurred. Or at least nobody has raised a cause celebre about them recurring.
TLDR: Old news. Maybe notso-old facts.
As I understand it, such devices have enough on-board voice recognition capabilities to recognize their wake phrases, but not enough to understand general speech: For that, they have to send the audio to cloud servers. And they can’t always be constantly sending an audio stream to the cloud, because that would use both too much bandwidth and too much cloud computing resources (just because it’s “in the cloud” doesn’t mean all that computation is free, or doesn’t need hardware to be running it). I suppose that they could be sending and processing randomly-sampled selections of non-triggered speech, but they can’t be doing it anywhere close to all of the time.
We conducted our own homebrew experiment as we have a few of these devices. For at least a couple weeks, all of us in the household would randomly talk about how much we wanted a parakeet, where to buy a parakeet, where to buy parakeet supplies, etc.
We have no pets, have no pet store discount cards, no pet supply orders on record, no (that we can recall) pet supply related searches…
And we waited to see if we ever got any sort of pet/pet supply ads popping up. None appeared to date. This was several months ago.
I seem to recall the belief in this was quite widespread. What a very interesting little case study in … something.
So is QAnon. People don’t understand technology or probability very well.
Cool experiment. Nice thinking.
I seriously doubt any of them are purposefully doing that.
I seriously doubt that, too. I was speaking of what is theoretically possible, if one were being fully paranoid about it.
It may be more sinister than the device listening in on conversations. It may be that the internet is initiating the conversation, that you are getting suggestions from your browsing that induce you to want a plate of shrimp or a new car, and when the pump has been sufficiently primed, here comes the advertisement, and you think “Wow, shrimp does sound good” or “That really isn’t my father’s Oldsmobile”.
Considering that Oldsmobile had been defunct for over 16 years, I doubt the internet is initiating anything in that direction.
It’s possible that people offering up theoretical possibilities at the paranoid level (and without making it clear at the time that even they themselves don’t believe it) may be one of the reasons the OP or people like them think this is a bigger thing than it is.
I thought I made it quite clear that I was asking what became of the belief that this was happening. It was once commonly thought to be true. Then, other stuff took over the public discussion and this idea went away.
I did not say I thought it was true.
I… just ate a plate of shrimp. Hmm.
