Police seek Amazon Echo data in murder case • Hacker News Board

Re: Police seek Amazon Echo data in murder case

#61

Post by jasonkostempski » Tue, Dec 27, 2016, 5:44 PM UTC

tclancy wrote: post #33

cryptarch wrote: post #13

Earlier quoted context omitted.

It would cost about $30m a year if you tailor the system to flagging specific data for storage and don't naively store every moment (e.g. you scrap silent moments and use VBR encoding). Storing a year's worth of 96kbps audio costs 380GB. If you don't record silence and you assume the people around an Alexa are only speaking for at most 4 hours a day on average, that goes down to 76GB a year. So if you then assume 5m…

I really like the math here, but isn't this a bit pointless? The system wants to parse meaning from audio; storing just the text it parsed is a lot smaller. Store just the text and whatever machine learning score of how probable the text is correctly parsed and that sounds like something prosecutors would love to bring into court: "Please read this line and let's see what score you get . . . "

For improvements they'd store the raw input so that when a mistake happens they can manually try to figure out why the machine got it wrong (e.g. a hi-hat was hit while they were saying "deuce" so it sounded like "douche").

Re: Police seek Amazon Echo data in murder case

#62

Post by gordyf » Tue, Dec 27, 2016, 5:45 PM UTC

Flammy wrote: post #28

FrancoDiaz wrote: post #24

Earlier quoted context omitted.

Wait a minute, we'd already know if Alexa was recording everything by the network traffic that obviously has already been sniffed. We can assume that Amazon is storing every "Alex,..." query that you've ever done. So what is there to know?

We know they are streaming it to do all the NLP in the cloud. This is obvious from the computational costs alone of doing NLP. The important question here is "What happens to the audio data streamed to the cloud, a few seconds AFTER the query has been identified (or not identified) and the results sent back to the Alexa device?" The nieve assumption is "all audio data is deleted after processing". The reality is that…

No need for speculation: you can open the Alexa app and play back your own queries, audio and all. There's even a help section titled, "How do I delete Alexa voice recordings", which reads:

"You can delete specific voice recordings associated with your account by going to History in Settings in the Alexa App, drilling down for a specific entry, and then tapping the delete button. Or, you can delete all voice recordings associated with your account for each of your Alexa-enabled products, by selecting the applicable product at the Manage Your Content and Devices page at www.amazon.com/mycd or contacting customer service."

Re: Police seek Amazon Echo data in murder case

#63

Post by coding123 » Tue, Dec 27, 2016, 5:47 PM UTC

cryptarch wrote: post #13

bkbridge wrote: post #2

Alexa only waits for the wake word "Alexa." It would cost more money than God in hardware to store every thing Alexa ever heard. She saves you're requests. And shows you those request in the Alexa App. And you can also delete it if you wish. There is no data for the police to have, because beyond requests, there is no data. Unless someone knows more about this than Amazon is telling us?

It would cost about $30m a year if you tailor the system to flagging specific data for storage and don't naively store every moment (e.g. you scrap silent moments and use VBR encoding). Storing a year's worth of 96kbps audio costs 380GB. If you don't record silence and you assume the people around an Alexa are only speaking for at most 4 hours a day on average, that goes down to 76GB a year. So if you then assume 5m…

Actually you really only need to store about 1 month of data. If the police request something Amazon can lock the account from auto-deletion.

Re: Police seek Amazon Echo data in murder case

#64

Post by supergeek133 » Tue, Dec 27, 2016, 5:50 PM UTC

notlisted wrote: post #48

Alexa listens only to the words after the wake word. What's more it doesn't capture more than 10-12s past the wake word. If you have a device you can check this yourself (Alexa, One Mississippi, Two Mississippi…). It's actually hard to get it to capture this much. Take a breath or pause to long and capture stops sooner. The request of the police is nutty and indicative of not understanding the technology. Unless the…

It doesn't matter, people will still tinfoil hat that the thing is listening and uploading everything they say.

I had this exact same conversation with a number of people when the news came out about the Echo being in the Wynn Hotel in Vegas.

Re: Police seek Amazon Echo data in murder case

#65

Post by chrishacken » Tue, Dec 27, 2016, 5:52 PM UTC

cryptarch wrote: post #13

bkbridge wrote: post #2

Alexa only waits for the wake word "Alexa." It would cost more money than God in hardware to store every thing Alexa ever heard. She saves you're requests. And shows you those request in the Alexa App. And you can also delete it if you wish. There is no data for the police to have, because beyond requests, there is no data. Unless someone knows more about this than Amazon is telling us?

It would cost about $30m a year if you tailor the system to flagging specific data for storage and don't naively store every moment (e.g. you scrap silent moments and use VBR encoding). Storing a year's worth of 96kbps audio costs 380GB. If you don't record silence and you assume the people around an Alexa are only speaking for at most 4 hours a day on average, that goes down to 76GB a year. So if you then assume 5m…

Based on the costs of disks alone, their cost per PB is actually around $35k. They likely get a volume discount, so we can lower that estimate even more and say $25-30k. Bandwidth is essentially free even though they charge ridiculous amounts of money for it on their services. You can get a 10Gbps link for as low as $2,000 if you buy it on-net and in bulk, Amazon probably gets it even cheaper. So ~3PB/month for $1,000/month.

Re: Police seek Amazon Echo data in murder case

#66

Post by JamieF1 » Tue, Dec 27, 2016, 5:52 PM UTC

smileysteve wrote: post #52

guscost wrote: post #32

This brings up an interesting question. If I say "Alexa, call the police!" or "Alexa, I need an ambulance!" will anything happen? It doesn't officially say it can be used to request emergency response and I'm scared to try it. I wonder if they've received a lot of these requests and/or if they have an emergency team on standby.

This would be of interest for the elderly, "help, I can't get up" with an omnidirectional microphone instead of a fragile, not waterproof, awkward beeper that you carry everywhere

I'm going to look in to this and develop a skill that allows someone to ask for help. Unless it has already been done,will check it out.

Re: Police seek Amazon Echo data in murder case

#67

Post by codingdave » Tue, Dec 27, 2016, 5:53 PM UTC

notlisted wrote: post #48

Alexa listens only to the words after the wake word. What's more it doesn't capture more than 10-12s past the wake word. If you have a device you can check this yourself (Alexa, One Mississippi, Two Mississippi…). It's actually hard to get it to capture this much. Take a breath or pause to long and capture stops sooner. The request of the police is nutty and indicative of not understanding the technology. Unless the…

I'm not sure you and I have the same definition of "listening". It must listen to hear the wake word, and likewise must process incoming sounds to do so. If you know of a way to do that without storing the incoming sounds, even if just briefly in temporary memory, I'm all ears...

Re: Police seek Amazon Echo data in murder case

#68

Post by mnm1 » Tue, Dec 27, 2016, 5:55 PM UTC

alexwebb2 wrote: post #8

CodeWriter23 wrote: post #7

From the privacy perspective, IMO, once you put a device in your home that you know is always listening, and is forwarding an unknown amount of audio from your home to a third party, at that point, you have surrendered your right to privacy. Edit: not just my opinion, it's a matter of law. (thanks @mnm1) https://en.wikipedia.org/wiki/Third-party_doctrine

No, you haven't, any more than when you buy a cell phone with a camera and microphone built in. People still have rights regardless of the products they buy.

Actually, according to the third party doctrine, you have. The difference between this and a cellphone is that you voluntarily give up this information whereas a cellphone does not transmit such data voluntarily. There are surveillance laws and other problems if a phone is recording you unknowingly. With Alexa, the recordings are submitted voluntarily. I'd be shocked if the police don't get their hands on this data. It's not like Amazon can claim, like Apple, that they can't themselves access it because it's encrypted.

'The third-party doctrine is a United States legal theory that holds that people who voluntarily give information to third parties—such as banks, phone companies, internet service providers (ISPs), and e-mail servers—have "no reasonable expectation of privacy."'

https://en.wikipedia.org/wiki/Third-party_doctrine

Re: Police seek Amazon Echo data in murder case

#69

Post by politician » Tue, Dec 27, 2016, 5:58 PM UTC

The device doesn't start listening when you say the wake word, the device is always listening. It's only the UX that responds when the presence of the wake word is detected. This means that the device is continuously monitoring the environment for the presence of the wake word we know about ("Alexa").

Can the device monitor for other words that trigger other UX sequences? Of course it can!

Can Amazon deliver an OTA update to this network connected device to change behavior after it leaves the factory? There is no reason to believe that they cannot.

So, no, I don't believe the police are wrong here. In fact, I would be in favor of law enforcement having the ability to "tap" these network-attached microphone appliances with a warrant. Then, perhaps, people will begin to understand what they've really bought.

EDIT: Passive monitoring ("listening") is not the same thing as storage ("recording"). In some cases below, these are being conflated.

Re: Police seek Amazon Echo data in murder case

#70

Post by roflchoppa » Tue, Dec 27, 2016, 6:01 PM UTC

Interesting that the police is going after this electronic speech data, reminds me of the instance where a parrot witnessed a murder that was posted on this site couple of months ago, in that case the parrot was not able to testify...

Also interesting that the lawyer is going for the reasonable expectation of privacy regarding that data, if that holds, then it could allow them to argue due protections from the 4th amendment.

But let's ask the question here; is data regarding usage of utilities private? It seems that something that you installed for water usage, should be. Could this data not have been acquired from the water department?