Amazon goes to court to stop US murder cops turning Echoes into Big Brother house spies Police in the US believe an Amazon Echo overheard the murder of a bloke found dead floating in a hot tub. All the cops want is a copy of any audio recorded by the personal assistant, conveniently stored in Amazon's cloud. One small problem: Amazon is still refusing to hand over any recordings, and has gone to court to shoo …
"Amazon is keen to ensure that the police can't use its devices as perpetual bugs in the homes of their owners. If legal precedent is established for that, sales will go off the side of a cliff."
Really? So many people who are happy enough to let Amazon use its devices as perpetual bugs in their home would be so terribly upset if the police could (but unlike Amazon almost never would)?
But you need to remember that Amazon (its subsidiaries, delegates, partners, etc) are only interested in making money based on what you give them access to (like all the audio in your home), meanwhile those nasty government agents are trying to do nefarious things like uphold the law.
I wonder whether the Amazon device detected and recognized the murder in progress then offered to expedite ship weapons, body bags or cleaning supplies. (Before the cops arrive with your Prime membership)
"But you need to remember..."
How about we rephrase that as "Amazon are only interested in selling you things while those nasty government agents are trying to do nefarious things like potentially single you out for prosecution based on something you might have said they decide they didn't like hearing, where for all you know they enjoy complete freedom in deciding what that might be, at any point now or in the future". Sounds a bit more honest to me. Further reading, endorsed by Schneier (and this is actually orders of magnitude worse than that, because you'd never know when they listen...).
Your wish is CNET's command...
At the end of the article:
https://www.scribd.com/document/335167572/Warrant-served-to-Amazon-re-Echo-data#from_embed
https://www.scribd.com/document/335167614/Warrant-served-to-Amazon-re-Echo-data#from_embed
IANAL and all that, but I do not see how Amazon's 1st amendment claims fit in all this. Unless someone really meant to say 4th. And even then, the device belonged to the victim, so that route is dubious as well. Someone at Amazon is getting desperate to keep the Big Brother tag off their listening posts.
"Someone at Amazon is getting desperate to keep the Big Brother tag off their listening posts."
Yep. In theory, Only speech/sounds listened to after the wake-up phrase is sent to Amazon and stored but is that really the case. And when does Amazon know to stop storing, assuming they stop? This may be the real reason Amazon are reluctant. The warrant asked for all stored interactions over a 48 hour period. People might well be shocked if it turns out more is stored than they assumed or were lead to believe.
But this is something else entirely.
First, it's not warrantless. Second, it's not speculative - an actual crime was committed, and a proper one at that, not merely something that offends the government's misguided sense of entitlement. Third, honouring a warrant to assist with a murder investigation will not magically set a precedent whereby everyone's privacy can be legally violated ... other than by Amazon who were doing so in the first place, of course.
Given the circumstances, Amazon's response is frankly bizarre.
Third, honouring a warrant to assist with a murder investigation will not magically set a precedent whereby everyone's privacy can be legally violated ...
Actually, here in the States, it can. The US law and enforcement, etc. has much basis on precedent. On the upside, if the court forces Amazon to do so, then the only place this precedent can be used is Alabama. Unless, of course, it gets appealed upwards to the Federal level. Then all bets are off.
>First, it's not warrantless.
I think the point point is: the police's surveillance isn't warrant-less, but Amazon's is.
The whole thing looks fishy to me. It would be easy to comply with this request and say, "see, we keep nothing" which would make them look good. If on the other hand, they are snooping more than they should, then things would get ugly and no-one would trust them, or Siri or Tay....
Of course, if he yelled, "Alexa, call the police!" and Amazon did nothing, that would look bad, even if it is entirely excusable.
This type of case is never a one-off but rather its used to set a precedent. In this case the Echo is being described as a recording device which it blatantly is not -- its a voice recognition terminal. Often people who appear on the surface to be quite smart are really dumb, incapable of learning anything. In this case they have got it into their heads that this is an audio recording device. It doesn't matter that the tear down of an Echo shows a lot of signal processing going on inside the unit. It doesn't even matter that what Amazon says it does because we're now so screwed up as a society we think that any time something doesn't fit our preconceived idea of what should be going on then its obviously a conspiracy, cover-up or some such..
If the cops had their way they'd force Amazon into making it a surveillance device which would pretty much kill off the market for this technology.
...and remember, that this is not America but New Amerika, the land of the not too thinly disguised jackboot. You don't have to have read much 'fake news' to realize that we're in the fight of our lives for the Bill of Rights.
@martinusher
"In this case the Echo is being described as a recording device which it blatantly is not -- its a voice recognition terminal."
I think that's one of the points of this case. If it isn't a recording device, Amazon can simply comply by saying we have nothing, it's not a recording device. If it is a recording device, Amazon will need to supply the recordings. At the moment, it's looking rather like the latter (hence Amazon refusing the request and fighting it). At that point, what happens to peoples buying habits and does the maket fall out of this type of device?
If there were an old-fashioned voice-activated tape recorder in the room, there would be no question about the police having the authority to play back the tape. The legal distinction here is that the microphone is in the room, but the recorder is elsewhere, but again if it were a physical device I doubt that there would be any question it could be seized as evidence.
What's bizarre about Amazon's response is that it is on the one hand saying that Alexa is not the equivalent of a tape recorder in the room but simultaneously implying it is by refusing to hand over data they claim not to have. It sounds to me more akin to the government's argument that it isn't engaged in mass surveillance because, although it is engaged in the mass collection of data, it only actually examines the data related to specific subjects of enquiry.
Also, if I were Amazon, I might be contemplating what benefit I might get in future by the more continuous processing of sound - identifying the number of people in the room, the mood of the conversation, possibly the topics - and feeding that back into my marketing algorithm. While it's not in their interests for Alexa to become known as the spy-in-the-room, it's also not in their interests to make definitive statements about what Alexa may or may not be listening to in future.
I don't think the police is really interested in the data gathered in this particular case. This is only a good excuse -OMG! A murder!- for setting a precedent, in a similar way to what the FBI tried in the San Bernardino attack.
Expect many more cases like this, where LEAs and TLAs -and not only in the USA- will try to use the most heinous crimes available to set legal precedents, with the hidden goal of obtaining the right to access any of these Big Brother electronic devices at their whim.
I personally wouldn't touch Siri, Alexa, Echo or Cortana with a bargepole, and most well informed people will be doing the same in a few years, I hope.
EDIT: After posting this comment, I noticed Mark 85 wrote just before a comment along the same lines, probably with a difference of seconds. Damn! :-)
I suspect that if this sort of discovery proceeded and the results (with regards to how much information is collected and stored) wound up in court, it would make people feel very uncomfortable having these things around. And from a legal standpoint, it would be nice if they were only used to solve murders but these precedents never, ever just stop there. Six months later they'd be combing through your digital records to resolve arguments over lunch money or other sundry trivial matters.
But the point is still valid - if they collected evidence of a crime and it is accessible to Amazon, they should hand it over. The iPhone case was very different, it was not accessible to Apple and they were being asked to assist in breaking the device's security which has much wider implications as it changes what other's get.
Sure, this might become a snooper's gold mine, but following the court case people should know the truth. If Amazon are recording all of this and storing it and that creeps you out - don't get one. Sadly this is likely the reason for the fight - money, not rights.
The iPhone case was extremely interesting in that it revolved around the ability of a government or govermental organisation to allow a company to break the law. Effectively, the argument was that Apply should be allowed to break hacking laws.
In this case, if Apple have recordings, the bigger question is about people being aware their audio is being recorded and kept, plus what Amazon do with it, hence the purpose of recording and keeping it. Be interesting to see what the legal agreement with these devices says as well, although knowing Amazon (and other big corporations), it probably allows them to do almost anything.
The idea of requiring a warrant is explicitly about not setting precedents, otherwise there would be no point in having warrants in the first place.
On the other hand, I agree that the principle does not always align with reality, sadly, especially in post-911 America, where even the most banal trivia is often magically reclassified as "a matter of national security". The US justice system seems to have been rewritten in Java, where the new motto is "Read once, read anywhere".
If, for those expressly wanting a "digital witness" in their home as part of a security setup, you could elect to have any reasonable evidence made available in the event of a suspected crime, this could be useful, a bit like the dash cam in your car.
I used (circa 2000) to want the security company monitoring my home alarm to be able to view internal cameras so that if they got an activation and I was out, I didn't have to return home or get a key holder before the police were called. Of course their link would have been activated only by a switch triggered by alarm conditions, not available to them 24/7, but it never came around. Now that tech has changed, this might be an option.
That said, I appreciate the concerns about mission creep, abuse etc and would not rush to be first in the queue until sensible people reviewed and approved the privacy terms.
I think there are two aspects to this that differ greatly from the case in hand.
Firstly, it's about openness on the devices capability, what is actually being performed. In this case, there's doubt about what Amazon are keeping and what they're doing with whatever they keep. In your case, you would know and be clear what was happening.
Secondly, in your case, you are giving express permission and defining boundaries to the actions and activities. In the other case, no express permission or boundaries are really defined.
"We need that recording to solve this murder."
When the Police are successful, it becomes "We need that recording to solve this attempted murder."
And then "We need that recording to solve this assault.", followed by "We need that recording to solve this robbery."
And before long it becomes "We need that recording because we think these guys are up to no good."
Scope creep is a watch word in Policing circles...
According to Amazon's court documents,they've offered to provide materials to the court if the court says "heightened" need has been demonstrated. The "heightened standard" was required to obtain a list of book purchases by a White House intern in a rather well-known case of the 1990s (as the books one purchases are regarded as 'expressive content' and thus protected), and in dozens of other cases cited. In other cases, Amazon purchases have been covered by this. And surely what someone says in the privacy of their own home to a box of bits is expressive content (and therefore so is the response) unless the police have a prior warrant to record it - so all the police have to do is show heightened need, and not try to get the precedent, and the Amazon "no you cant have this" goes away.
If the police show "heightened" need, Amazon will turn it over, and no new precedent has been set. Otherwise this would seem very much like they are trying to set a precedent. If i was being unkind I'd say the precedent they want is that of using a search warrant as a means of obtaining a retrospective surveillance / wiretap warrant, but I'm sure no government agency would ever even dream of doing something that, any more than they would allow anti-terror legislation to be used by local councils, not just the security/police services.
The only reason I can think of for Amazon to refuse this is if they actually record AND STORE everything it hears 24x7, and save it indefinitely. If they're really doing that, I can see why they wouldn't make to make that known by producing this evidence.
If they are doing what you would think is reasonable, and tossing anything it hears that isn't intended for it, then Amazon could say "sorry coppers, we aren't selling an always-on spy device so we don't have the data you're requesting!"
It is one thing to SUSPECT that companies like Amazon and Google are saving everything a device like this hears forever, quite another to have it essentially CONFIRMED! Hopefully this case will get lawmakers in a tizzy on both sides - the ones who want unlimited police powers will want laws that force Amazon to comply, those who don't will try to instead fix the problem by requiring companies that make devices that act as a 24x7 bug note it prominently on the packaging!
Maybe this would be a place for cooperation - make a law that says if you sell an always-on recording device that you must give up the data if presented with a legal search warrant, AND you must prominently label the device with a notice that it saves everything it hears. Let's see how the Amazon/Google business model for these assistants holds up with that labeling forced on them!
I want to know why this information is being sent even if the device is not triggered.
I don't understand why the alert phrase is not identified locally to switch on the recording. I mean, recognizing one of three words to activate the device is not particularly difficult, and providing it worked as advertised, would prevent Amazon recording things other than what's intended.
In fact, I would prefer that a majority of the voice recognition was done locally, so there would be a chance that they could do something useful even when not connected to cloud services. Make them use my NASor music server to find media, use a local calender, and only go out to the 'net when it could not satisfy a request locally.
But I suspect that one of the primary reasons these things exist is to get people used to an always connected house.
Nice wording on Amazon's part about the storage of chats.
Tip: To delete individual recordings, select Delete voice recordings. This removes the audio file streamed to the Cloud, as well as the Home screen cards related to that interaction.
No point on saying where it's removed from because that might confuse the issue, right Jeff? I assume it's removed from the playlist like interface and not actually removed from the Cloud's data store.
to allow police see any recordings in the event I'm murdered. I can understand the privacy issue if they just tried a general blanet request to see what I've been saying, but FFS the bloke has been killed, the poilce need all possible evidence.
It's incredibly unlikely the device recorded anything but a murder investigation covers everything, 99% is probably irrelevant.
We have two echos in our house + google now on three phones and 2 tablets., i would not call them invasions of privacy, they only upload audio when the Echo/Alexa/Dot/Google is triggered. There is a reason these and similar services can only be triggered by a discreet set of words, the devices are optimised to listen for those words, subsequent audio is then uploaded for processing.
From what I know. I've done some development with Alexa and monitored network traffic between it.
Plus, despite what people may think of Google/Amazon/Big Corp, I seriously doubt they would be continuously gathering what is being said whilst their product description clearly states what it does.
Just using it it is clear that the trigger word is detected locally, they have difficulties with a a 5 year old kid trying to activate it but once triggered, both Alexa and Google Now have no trouble understanding the search term. I doubt either would do anything to damage their reputation by blatantly lying.
Alexa does occasionally get triggered falsely so there is certainly some audio uploaded that was not intented, however the Aleax app lets you see the history and remove things.
This post has been deleted by its author
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
