8 ms·

So essentially they asked for anonymised lists of phones (i.e. not even the actual numbers) that had been in the locations where the suspects had been, then cro

So essentially they asked for anonymised lists of phones (i.e. not even the actual numbers) that had been in the locations where the suspects had been, then cross-referenced these to find potential suspects, presumably finding some phones that exactly matched the reported movements of the suspects which they then made a second application to de-anonymise. Do people honestly think this is disproportionate for investigating a home invasion? It seems like they have been quite privacy conscious in my opinion. What would have been a better way to investigate this crime?

michaelt 8y ago
Do people honestly think this is disproportionate for investigating a home invasion? I think the concern is not Google sharing this data once they have it. The concern is Google having it, in the first place (in personally identifiable form). There's a lot of unclarity about which free and paid Google services and products are and aren't tracked. Who knows if their DNS resolver or their wifi-visibility-to-phone-location service or google-maps-used-without-logging-in would have come up in such a search warrant?
- bagacrap 8y ago
  Hold on, this is great justification for Google having this data (in addition to the real reason, which is to provide better location-based services). Due to the anonymization process, no one's privacy is unduly infringed upon, and investigators have a tool for catching criminals. Do you not see that last bit as a societal win?
- bepvte 8y ago
  We know, because its outlined in their privacy policy that they store dns logs for 24 hours, and store the location data of devices which allow the location history feature, and provide data to governments who send a proper search warrant with reasonable cause.
  - pubutil 8y ago
    My problem isn’t so much that Google is storing location data if people have agreed to allow it. The semantics of agreeing to enable location history are arguable, but by default Google’s location history feature is opt-out. I’d prefer it be not be that way, but that’s a debate for another time. What I do have a problem with is what this line from the article suggests: >An Associated Press investigation last year revealed that even with these settings disabled, Google continued to collect and store users' location data. Maybe I’m missing something from the TOS, but it seems to me that the opt-out process doesn’t do what it describes; so what is it for?
- ilovetux 8y ago
  > The concern is Google having it, in the first place This. I think it's the only responsible thing for the police to do. Since they know the data probably exists they have to look, but does this data need to exist? What we need is some limitations on what is collected and how long it is stored.
renholder 8y ago
>Do people honestly think this is disproportionate for investigating a home invasion? The scale of the search[0] is grossly disproportionate, to be sure. The population of the city in question was 60,797+[1] almost ten years ago. So, we have a wide-net around the house, a period of about six hours requested, and if that main road is a major artery for driving, then we can safely assume (if it's day-time time-frame) that at least a small-to-medium-size portion of the population would've been simply driving by. Now, for the St. Paul Food Market, they wanted 33 hours' worth of data and that is off of two main drags (it looks like). Given Google's overestimation of its ability to place a device at a location (noted in the article) and the potential of the predominance of the population to have driven in that area (let's say business day versus weekend); then, yeah, that does seem grossly disproportionate for investigating a robbery of a home invasion, and/or markets; especially considering that the markets weren't hit multiple times over that time-frame - they just wanted to see patterns of travel for everyone in the area over that period of time. >What would have been a better way to investigate this crime? The original police-work methods they were using, which they admitted in their court-filings as having actually led to the suspects? The confidential informant being chief amond those, yeah? Touting this method as being a more useful mechanism for investigation, when that was countermanded by the evidence of the events - as told in the article, is disingenuous - at best. [0] - https://img.apmcdn.org/170e8a92a0f2ddb41b4a415707a8e4db547af0f7/uncropped/720223-20190207-warrant-coverage-areas.jpg https://img.apmcdn.org/170e8a92a0f2ddb41b4a415707a8e4db547af... [1] - https://en.wikipedia.org/wiki/Eden_Prairie,_Minnesota https://en.wikipedia.org/wiki/Eden_Prairie,_Minnesota
- JoeSmithson 8y ago
  The additional search areas limit the potential for collateral intrusion though, because they are essentially going to be ANDed with the list from the main crime scene not ORed. Even if the whole town visited those markets, it could never be more intrusive. Also, the reasonably wide rectangle is precisely because the location data is not particularly accurate, I am not sure why you think this supports the idea that it should be smaller. I'm honestly shocked that people think that the police are overstepping the mark by requesting anonymised phone data from around the scene of a home invasion murder. Over the last few years I've watched the privacy debate on HN evolve from completely legitimate criticism of mass surveillance programs with insufficient transparency and oversight, to criticising a detective getting a warrant from a judge, that is specifically written to request anonymised data! While investigating a murder in his community! > The original police-work methods they were using, which they admitted in their court-filings as having actually led to the suspects? The confidential informant being chief amond those, yeah? The level of proof required in a criminal case is incredibly high, what's more police have a responsibility to find all assailants as well as test the veracity of witness reports and follow all reasonable lines of enquiry. "Witness is lying", "Someone stole my car" etc etc. All these will be presented as defences. They don't even need to be probable, because the prosecution needs to clear a very high bar of proof. This is why police collect as much evidence as they can and don't just stop when they think they know who did it. Ultimately the detective has a responsibility to the wife who's husband is dead - and to society in general - to make sure people face justice when they offend, and they can't half arse it. (Disclosure: I am a detective)
roenxi 8y ago
If your goal is to arrest someone for the crime, then this may be the best way; but if you optimise society for catching criminals you aren't optimising for prosperity, comfort or sustainability. The country with the greatest personal liberty, currently probably the US, is consistently a place that people would choose to live if they had a choice. This is a hint that moving away from personal liberty will end up with a society that is (1) free of crime and (2) not a desirable place to move to. I'm not really going to expand the argument; I hope it is self-evident why having your activities reviewed by a person who might choose to arrest you is going to make your life worse. Statistically and at the macro scale that is a bigger threat to you than a home invasion (even if you think you aren't committing any crimes - I hope you've read all the laws that apply to you, don't ever get into a grey area and the police never make a mistake!).
- lawlessone 8y ago
  >The country with the greatest personal liberty, currently probably the US What's their prison population again?
  - PavlovsCat 8y ago
    There's countries where the percentage is lower, yet it's basically illegal to criticize the government. More importantly, what does that change about the argument that arresting criminals is hardly the only thing to consider? Here's the same argument with enhanced contrast: If you threw Earth into the sun, all crimes would stop. If that's all we care about, everything else is just kicking the can down the road. But that would be a "crime" in and of itself, right? IMO the same could be said about mass surveillance, it's stripping humans of their rights to preserve their rights, burning the village to save it. Generally I can't help but notice that it's really not the size of the crime that seems to matter, but whether it's aligned with centers of power or not. We're aiming at a world where a tyrant can't be assassinated, and the defenseless can be murdered without consequence, with a bit of playing for time and lip service to principles here and there. > It is forbidden to kill; therefore all murderers are punished unless they kill in large numbers and to the sound of trumpets. -- Voltaire
gbrown 8y ago
We have no reason to trust either the sensitivity oe specificity of such an identification, and every reason to assume police and juries will think it's foolproof.
rocqua 8y ago
Replacing IDs by Psuedonyms does not really give anonymity. Consider correlation with CCTV cameras, or identification based on sleeping in a house. As an hypothetical, suppose there is data on two people, who were in the same house at 11:00 pm. Then, at 12:00 pm, one of those people moved over to a neighboring house for 1:00 hour, after which they went back. It looks a lot like two people who live together, where someone snuck out to spend the night with someone else. This is sensitive data, and not really anonymous.
jancsika 8y ago
> So essentially they asked for anonymised lists of phones (i.e. not even the actual numbers) that had been in the locations where the suspects had been, then cross-referenced these to find potential suspects, presumably finding some phones that exactly matched the reported movements of the suspects which they then made a second application to de-anonymise. In what conceivable way does that count as an "anonymised list?" Suppose Google happens to sell geodata for everyone in your area of town to a third-party. Are you reassured if they tell you, "It's ok, we removed your phone numbers first."

michaelt 8y ago

Do people honestly think this is disproportionate for investigating a home invasion? I think the concern is not Google sharing this data once they have it. The concern is Google having it, in the first place (in personally identifiable form). There's a lot of unclarity about which free and paid Google services and products are and aren't tracked. Who knows if their DNS resolver or their wifi-visibility-to-phone-location service or google-maps-used-without-logging-in would have come up in such a search warrant?

bagacrap 8y ago
Hold on, this is great justification for Google having this data (in addition to the real reason, which is to provide better location-based services). Due to the anonymization process, no one's privacy is unduly infringed upon, and investigators have a tool for catching criminals. Do you not see that last bit as a societal win?
bepvte 8y ago
We know, because its outlined in their privacy policy that they store dns logs for 24 hours, and store the location data of devices which allow the location history feature, and provide data to governments who send a proper search warrant with reasonable cause.
- pubutil 8y ago
  My problem isn’t so much that Google is storing location data if people have agreed to allow it. The semantics of agreeing to enable location history are arguable, but by default Google’s location history feature is opt-out. I’d prefer it be not be that way, but that’s a debate for another time. What I do have a problem with is what this line from the article suggests: >An Associated Press investigation last year revealed that even with these settings disabled, Google continued to collect and store users' location data. Maybe I’m missing something from the TOS, but it seems to me that the opt-out process doesn’t do what it describes; so what is it for?
ilovetux 8y ago
> The concern is Google having it, in the first place This. I think it's the only responsible thing for the police to do. Since they know the data probably exists they have to look, but does this data need to exist? What we need is some limitations on what is collected and how long it is stored.

renholder 8y ago

>Do people honestly think this is disproportionate for investigating a home invasion? The scale of the search[0] is grossly disproportionate, to be sure. The population of the city in question was 60,797+[1] almost ten years ago. So, we have a wide-net around the house, a period of about six hours requested, and if that main road is a major artery for driving, then we can safely assume (if it's day-time time-frame) that at least a small-to-medium-size portion of the population would've been simply driving by. Now, for the St. Paul Food Market, they wanted 33 hours' worth of data and that is off of two main drags (it looks like). Given Google's overestimation of its ability to place a device at a location (noted in the article) and the potential of the predominance of the population to have driven in that area (let's say business day versus weekend); then, yeah, that does seem grossly disproportionate for investigating a robbery of a home invasion, and/or markets; especially considering that the markets weren't hit multiple times over that time-frame - they just wanted to see patterns of travel for everyone in the area over that period of time. >What would have been a better way to investigate this crime? The original police-work methods they were using, which they admitted in their court-filings as having actually led to the suspects? The confidential informant being chief amond those, yeah? Touting this method as being a more useful mechanism for investigation, when that was countermanded by the evidence of the events - as told in the article, is disingenuous - at best. [0] - https://img.apmcdn.org/170e8a92a0f2ddb41b4a415707a8e4db547af0f7/uncropped/720223-20190207-warrant-coverage-areas.jpg https://img.apmcdn.org/170e8a92a0f2ddb41b4a415707a8e4db547af... [1] - https://en.wikipedia.org/wiki/Eden_Prairie,_Minnesota https://en.wikipedia.org/wiki/Eden_Prairie,_Minnesota

JoeSmithson 8y ago
The additional search areas limit the potential for collateral intrusion though, because they are essentially going to be ANDed with the list from the main crime scene not ORed. Even if the whole town visited those markets, it could never be more intrusive. Also, the reasonably wide rectangle is precisely because the location data is not particularly accurate, I am not sure why you think this supports the idea that it should be smaller. I'm honestly shocked that people think that the police are overstepping the mark by requesting anonymised phone data from around the scene of a home invasion murder. Over the last few years I've watched the privacy debate on HN evolve from completely legitimate criticism of mass surveillance programs with insufficient transparency and oversight, to criticising a detective getting a warrant from a judge, that is specifically written to request anonymised data! While investigating a murder in his community! > The original police-work methods they were using, which they admitted in their court-filings as having actually led to the suspects? The confidential informant being chief amond those, yeah? The level of proof required in a criminal case is incredibly high, what's more police have a responsibility to find all assailants as well as test the veracity of witness reports and follow all reasonable lines of enquiry. "Witness is lying", "Someone stole my car" etc etc. All these will be presented as defences. They don't even need to be probable, because the prosecution needs to clear a very high bar of proof. This is why police collect as much evidence as they can and don't just stop when they think they know who did it. Ultimately the detective has a responsibility to the wife who's husband is dead - and to society in general - to make sure people face justice when they offend, and they can't half arse it. (Disclosure: I am a detective)

roenxi 8y ago

If your goal is to arrest someone for the crime, then this may be the best way; but if you optimise society for catching criminals you aren't optimising for prosperity, comfort or sustainability. The country with the greatest personal liberty, currently probably the US, is consistently a place that people would choose to live if they had a choice. This is a hint that moving away from personal liberty will end up with a society that is (1) free of crime and (2) not a desirable place to move to. I'm not really going to expand the argument; I hope it is self-evident why having your activities reviewed by a person who might choose to arrest you is going to make your life worse. Statistically and at the macro scale that is a bigger threat to you than a home invasion (even if you think you aren't committing any crimes - I hope you've read all the laws that apply to you, don't ever get into a grey area and the police never make a mistake!).

lawlessone 8y ago
>The country with the greatest personal liberty, currently probably the US What's their prison population again?
- PavlovsCat 8y ago
  There's countries where the percentage is lower, yet it's basically illegal to criticize the government. More importantly, what does that change about the argument that arresting criminals is hardly the only thing to consider? Here's the same argument with enhanced contrast: If you threw Earth into the sun, all crimes would stop. If that's all we care about, everything else is just kicking the can down the road. But that would be a "crime" in and of itself, right? IMO the same could be said about mass surveillance, it's stripping humans of their rights to preserve their rights, burning the village to save it. Generally I can't help but notice that it's really not the size of the crime that seems to matter, but whether it's aligned with centers of power or not. We're aiming at a world where a tyrant can't be assassinated, and the defenseless can be murdered without consequence, with a bit of playing for time and lip service to principles here and there. > It is forbidden to kill; therefore all murderers are punished unless they kill in large numbers and to the sound of trumpets. -- Voltaire

gbrown 8y ago

We have no reason to trust either the sensitivity oe specificity of such an identification, and every reason to assume police and juries will think it's foolproof.

rocqua 8y ago

Replacing IDs by Psuedonyms does not really give anonymity. Consider correlation with CCTV cameras, or identification based on sleeping in a house. As an hypothetical, suppose there is data on two people, who were in the same house at 11:00 pm. Then, at 12:00 pm, one of those people moved over to a neighboring house for 1:00 hour, after which they went back. It looks a lot like two people who live together, where someone snuck out to spend the night with someone else. This is sensitive data, and not really anonymous.

jancsika 8y ago

> So essentially they asked for anonymised lists of phones (i.e. not even the actual numbers) that had been in the locations where the suspects had been, then cross-referenced these to find potential suspects, presumably finding some phones that exactly matched the reported movements of the suspects which they then made a second application to de-anonymise. In what conceivable way does that count as an "anonymised list?" Suppose Google happens to sell geodata for everyone in your area of town to a third-party. Are you reassured if they tell you, "It's ok, we removed your phone numbers first."