> this type of thing is certainly within the realm of possibility
In your scenario, Bezos is an evil worldwide overlord who murders non-supporters casually. And yet somehow the real problem is an Echo on the kitchen counter.
Just... what the fuck?
“Realm of possibility” means something very different to you than to me.
Overlords who casually murder non-supporters is not far-fetched by any stretch, such regimes have risen dozens of times just in the last 100 years. These overlords haven't quite reached "worldwide" status yet, but use your imagination.
A device that instantaneously and irrevocably transmits anything spoken to it to HQ, and then automatically transcribes it and makes it available for analysis, is absolutely a dangerous tool for censorship and repression. The fact that it looks nice on your counter and also spits out weather reports doesn't automatically invalidate those risks.
These BigCos already compose complex, fairly accurate estimations of our interests and beliefs based on the data they read out of our emails, search history, and yes, voice commands. This data is used to make advertising decisions, but there's no reason that the same data can't be applied to "pre-moderate" the opinions of those who've expressed "dangerous views" in private, or worse. It just hasn't been used that way yet (to our knowledge, though it wouldn't surprise me to learn that Facebook or Google are already punishing/down-ranking accounts that express sentiments they dislike).
Inviting these companies into your house to sit in every room and listen for "wake words", and then send up snippets of audio to HQ, should not be dismissed as a casual thing, even if the tradeoff is considered acceptable by some.
This is absurd. When Bezos becomes the dark lord you imagine, just unplug your Echo and throw it away. Problem solved. In the meantime, your use or non-use of an Echo seems to have no meaningful impact on the rise of dark lord Bezos.
Also you well know that the Echo is not transmitting everything it hears to HQ. It transmits only when invoked, which can be spurious but is not constant. Same for Apple and Google equivalents.
I don’t think Bezos = Sauron... but this is just wrong:
This is absurd. When Bezos becomes the dark lord you imagine, just unplug your Echo and throw it away. Problem solved.
How much data will have been collected and stored by then? Let’s just put Bezos aside for a moment, and think about foreign and domestic intelligence orgs which will get their fingers into this. Let’s remember a bit of WWII, when harmless census data empowered the Nazis to track down people with Jewish ancestry.
It’s not about what you think Bezos or today’s govt will do with the information, but what the corporate raiders and govts will do with it for the rest of your life.
> How much data will have been collected and stored by then?
It doesn’t matter. If the concern is that Bezos will flip the echo to be a political surveillance device, there is an immediate fix. Yes, the existing cache of data may still be around, but that wasn’t the concern posed by the original poster. The concern was specifically that after Bezos becomes the dark lord, an “innocent” question could result in politically motivated murder.
If your concern with the echo is that you don’t want anything recording anything in your house, that is reasonable and you should not get an echo. Of course, you should probably also not carry a phone.
> Let’s remember a bit of WWII, when harmless census data empowered the Nazis to track down people with Jewish ancestry.
And yet we still do the census because the value is high and the concern low. The fact that a thing can be abused does not mean it should not be used.
It doesn’t matter. If the concern is that Bezos will flip the echo to be a political surveillance device, there is an immediate fix. Yes, the existing cache of data may still be around, but that wasn’t the concern posed by the original poster.
Well it’s my concern, and I suspect others have it too. I’m not trying to weigh in on the side of “Bezos is evil” by any means, I’m just adding a dimension that seemed to be neglected.
I totally understand that. It’s just not really the scenario proposed.
Personally I’m not very concerned about the Echo, which is why I own it. But I do understand the reasonable concern that records could be subpoenaed or the data feed could be tapped.
It transmits everything after the wake word. My example of the Echo being used to monitor the political education of children is very plausible. It depends on a child using the wake word and asking the Echo an innocent question without understanding the broader implications, e.g., that something bad can happen if you let some big powerful group know that you're asking questions that they don't want people to ask.
Last I checked, the Echo had 3 potential wake words: "Alexa", "Echo", and "Amazon". I don't use "Alexa" because I knew a kid named Alexa growing up and I didn't like her. I don't want to remember her every time I talk to the machine. I use "Echo" but when I'm listening to audiobooks or podcasts and the word "Echo" comes up, the thing activates, and maybe I _don't want_ Amazon to know what podcasts or books I listen to. It also happens on words that sound kinda-sorta like "Echo". Random "accidental" invocation and the loss of control over the listener then becomes a real thing, especially with a wake word like "Amazon".
All they'd have to do is say "We've found 'Amazon' is the best wake word and 90% of our people use it so it's the default now", and suddenly, they'd be hearing a lot of casual conversations and opinions about what people think of this 'Amazon'. Now imagine that this really does become a tool for political oppression and they add a few "silent wake words" that don't make the ring light up.
I have 3 Echos. My kids ask it questions all the time. If I were to teach my children some politically incorrect belief or some religious tenet that counters the dominant religious tenets of the day, and they ask the Echo for more information about it, someone at Amazon may hear that, and a computer at Amazon would certainly process and analyze it.
Some day, the result of that analysis may be flagging for further investigation or addition to dissident watchlists. They may even apply this to archived recordings, all of which you yourself can play back (for now).
Imagine someone from the other side of the aisle guessed $Prominent_Politician's Amazon password. What would happen if there was a voice recording of that politician asking something that doesn't look so good? Would the attacker just say "Oh boy I better not share this, that wouldn't be very nice?"
In this age when people like Jo Cox and Steve Scalise are targeted by radical citizen-assassins, how can we pretend there is no risk? How can we pretend like this data is somehow intrinsically impregnable to technically sophisticated political extremists? And what happens when a demagogue mainstreams that extremist perspective, and writes laws that require Google or Amazon to submit these recordings for the good of the people?
I'm not trying to be alarmist, and I obviously think that the tradeoff is not horrible right now since I have 3 of these spy devices (and a Nest cam), but we have to understand that these are real things, real possibilities. It is very naive to presume that this couldn't or wouldn't happen (or even that it doesn't on some scale -- I'm sure a subpoena or NSL could get this data).
> My example of the Echo being used to monitor the political education of children is very plausible.
“Plausible” does not mean “not literally impossible”. It means something is reasonably possible. Your scenario is not reasonably possible. It is exceedingly unlikely. It is possible, perhaps plausible, that the government could tap the echo data feed, but that’s really a much different scenario than the active monitoring of children’s political education that you imagine.
You clearly don’t believe in the plausibility of these scenarios or you would not have three Echos in your house.
>“Plausible” does not mean “not literally impossible”. It means something is reasonably possible. Your scenario is not reasonably possible. It is exceedingly unlikely.
"Plausible" means it reasonably could happen, yes. It doesn't necessarily have to mean that it's expected to happen in the immediate term (unless that's the pre-existing context to the discussion).
I don't believe conditions are such that it will happen right now. But that doesn't put it into the realm of mere theory, the things that make you say "I mean, sure, that's possible" behind rolling eyes.
Given the correct political conditions, I would venture that these eventualities are not just plausible, but downright expected. Maybe not the black bag, but certainly monitoring, penalization, enforced re-education (which portends black bagging).
>You clearly don’t believe in the plausibility of these scenarios or you would not have three Echos in your house.
Most of us do a lot of things that we know we really shouldn't do out of convenience, habit, etc.
I don't think it's likely that I'd get picked up for anything that my kids end up telling my Echo, at least not at this juncture, which is why I still have them.
If my religious or political beliefs were more fringe, or if the ground continues to crumble such that they become fringe by virtue of the cliff of acceptable discourse growing ever-shorter, then I would certainly consider it unsafe to let the children have access to the Echo. We monitor it now, but it's in a place where they can talk to it if they want. Wouldn't be that way if I thought we had something to worry about.
It's true that the archived recordings could also be used against us, but I think that we're moderate enough that there'd be too many other more worrisome targets to pursue first. Of course, you never really know.
Should things get this way, I would probably still keep one of the listeners hooked up in some room that is rarely used for conversation and feed it information that the bosses would be happy about. Maybe I would say "Echo, please tell me about all of Jeff Bezos's accomplishments."
In a more realistic climate where Bezos doesn't cross over into overt dictatorship, it may be smart to do this alternately with various political figures, so that I can cherry-pick proof of early admiration for whichever political figure wins and implements the dystopian future. This would be much smarter than just unplugging the device and having the line go dead, especially if they correlate the date the feed went dead with a specific political event.
I disagree? I don't think things will deteriorate quickly enough that it will be an issue on the current-gen hardware, but we're discussing the capabilities here, and why people may be skittish about giving these companies an unconstrained inlet into their homes. Technical plausibility does not necessarily imply there's a likelihood of deployment, though I think there would be if the right political circumstances came about. We're talking about the abilities of the devices and the ease of misuse, not whether Jeff Bezos fancies himself a future evil emperor or not -- only Jeff Bezos knows that!
All existing Echo models, afaik, are easily capable of being deployed and utilized for these purposes. I don't know whether a remote firmware update can change the wake words that can activate the device or not, but other than that, there is no good reason why this wouldn't happen other than just nobody has done it. From a technical perspective, this kind of use is completely plausible.
Whether or not the political will to exploit this capacity en masse will arise or not is a different matter, and I'm not really betting one way or the other on that.
> this type of thing is certainly within the realm of possibility
That's the same kind of thinking that makes people so scared of being victimized by violent crime even though things are safer now than they ever have been.
It's possible, but it's also possible that all of the carbon dioxide in the room randomly collects around your head and you suffocate.
In what ways? How do you think people will be harmed?