Apparently Google provided police with IP addresses, but not names [1]. Assuming that Google did not provide any other account details, the usage of a VPN service or Tor might be able to protect users in similar cases.
Or just use a sensible search engine like Kagi, which doesn't associate search queries with users (it doesn't need to; it's not in the business of selling users to ad brokers).
It's weird how easily we've acclimatized to creepy tech-company spying. For Google to snoop on your searches, record them, and permanently archive them for retrieval years in the future was a very creepy and ignominious business model, and it's embarrassing how readily we all went along with it, for the convenience.
I just went looking for my Google Search history (https://myactivity.google.com/product/search) and it claims "Not saving" but I know for a fact searches I'm making on my mac auto-complete on my pc if I use the "Google" container in Firefox.
It says the same "No activity" for YouTube but I can clearly go on YouTube watch history right now and see every video I've played. That's extremely fishy.
If I go to youtube activity controls and click "turn off" then my history stops being saved (checked by watching a video) then when I click "turn on" then the history starts working again.
The YouTube one could be if you have a old/brand account with a name that's different from your Google account. It should be available in the account switcher.
I'm not sure I buy the dragnet scare used on this. After a house is burned down, they are likely to ask everyone around if they saw or heard anything. Heck, they are likely to ask the library in town if anyone had checked out a book that outlined how to make a firebomb or similar. Ask local stores if they saw anyone buy fuel that could be used as an accelerant. Expectations are a no in all cases for questions like this, but they still ask. In this case, they asked, "any chance anyone has been looking at this particular address?"
It would be one thing if they asked for general access to every search that had been done in town. But, wasn't this a case of asking if they searched on a specific address, and it turning up a silly low number of results? And then they were able to find other leads and data?
Or is there a more sinister vibe that I'm just not seeing here?
It can stifle speech and knowledge, whether intentionally or unintentionally, the consequences are the same.
People are less likely to look up, or say, things in private that can get retroactively interpreted in the worst possible way by state actors who are looking for an easy arrest and/or conviction.
In the past, it was nearly impossible to round up everyone based on what they looked up in an encyclopedia, or on a map, in private. Your pursuit of knowledge wasn't limited by the very real threat of "Can just having curiosity about something land me in prison for terrorism/murder/arson/etc?"
It borders on thoughtcrime and feels like a fishing expedition to me.
> After a house is burned down, they are likely to ask everyone around if they saw or heard anything
To continue this analogy, simply asking is different than getting a warrant/subpoena for everyone in a neighborhood to search through records of their past speech to see if they said anything they deem suspicious.
If you had been yelling that you would be happy that someone's house burned down, I suspect you would be a suspect in the event that said house burned down. Does that mean we are stifling free speech that hopes harm on people?
I can see some concern over this, but just not in this case. If it was, "all people that searched pyromancy" and that included people that were looking up fireworks, for example. Or if it was all people that looked at any house on redfin in that state.
This was a fairly targeted search, all told. "Anyone been asking around about the victim?"
Rulings aren't only about the case. Precedent is important. Loosing rulings here can lead to expanded power and suppression of speech for fear of backlash.
Saying "checking for searches of an address in two weeks" is materially different than "searched for X suspected terrorist organization in four years".
I don't know that I follow your example. Searching a specific address is a fairly rare thing to do, all told. Such that I suspect you have a tighter target for that than you do any organization search.
Libraries haven't really been that big on privacy lately. They did fight against the government collecting our reading lists but when they lost that battle it feels like they just gave up. Now they encourage people to use 3rd party DRM filled apps that collect personal information including people's reading lists so that those apps can use that data to (among other things) push ads at them. That's the kind of thing a librarian would have strongly opposed a decade or two ago, but today they push those apps on adults and children without even mentioning the privacy issues.
That is, there is a good chance something like this request would be honored by any library. "Any chance you saw a person following the victim out of the building?".
> After a house is burned down, they are likely to ask everyone around if they saw or heard anything. Heck, they are likely to ask the library in town if anyone had checked out a book that outlined how to make a firebomb or similar. Ask local stores if they saw anyone buy fuel that could be used as an accelerant. Expectations are a no in all cases for questions like this, but they still ask. In this case, they asked, "any chance anyone has been looking at this particular address?"
When police ask for information in person, the people being asked know about it. When police instead go to Google with a dragnet warrant, the people included in the search never hear about it.
But there's a much more fundamental problem, a problem of constitutional rights. A warrant is based on probable cause to believe that someone commmitted a crime. The Fourth Amendment expects a flow like this: Police get a report about a crime. Police investigate the crime scene. Police go in person asking about suspects. With this preliminary evidence, the police have probable cause to pick suspects and get warrants to search the suspects.
A keyword search warrant is a general warrant. Police get a keyword search warrant. They search people who aren't suspects. Even if Google does the searching and decides which people to recommend to the police, Google does the searches because the government forces Google to search in the first place. The police use the information to get probable cause to pick suspects. This is backward: the searching comes before the probable cause. The Fourth Amendment breaks in this model.
If you think the warrant wasn't requested well, you kind of have a point. If the police simply had the "you know, I bet the arsonist would have used google maps to get directions to the place they burned down," that seems legit to ask for the information from Google. They would have already scanned phone records, and that seems... fine?
I could see arguments that they need to scope down the time of request for that warrant, as well. But flat out disallowing it feels just as wrong as opening it up for no process.
They may ask, and you may plead the fifth, and I might decline to talk to the police on principle.
Executing a general warrant compels everyone to “answer the question” in your analogy, which is clearly very different. You keep saying “ask” but here they seize.
That doesn't really change much, though? For one, they couldn't compel you, even with a warrant, to testify against yourself. They can compel you to provide information against others, though. If we are against that entire concept, that is one thing. But that is not the argument here, is it?
All someone would need to do to frame another for a crime like that is to hack their WiFi and set up a small device to run a very specific google search for a period? Hopefully they found more evidence than just the search queries.
And if this was the only evidence in the case, you would have a point? They did, in fact, find more evidence. Is why the defense took a long shot approach to try and break the first link in the chain.
Leaving aside that state supreme court decisions are overturned often, case law tries to be narrowly applied where it can. And this specific case is fairly narrow.
> The case is People v. Seymour, which involved a tragic home arson that killed several people. Police didn’t have a suspect, so they used a keyword warrant to ask Google for identifying information on anyone and everyone who searched for variations on the home’s street address in the two weeks prior to the arson.
So the warrant required producing information about a small number of Google users.
> Like geofence warrants, keyword warrants cast a dragnet that require a provider to search its entire reserve of user data—in this case, queries by one billion Google users.
...but they talk about the total number of people's records Google had to look at to find the small number that satisfied the warrant. I'm not sure why they are bringing that up.
As far as I know that's never been relevant when it comes to the 4th amendment. It's the result set that matters, not how many records you had to look at to find that set which can vary greatly depending on just how how you happen to organize your records.
E.g., suppose that two rental cars were involved in a crime, one from Hertz and one from Avis. Police know the license numbers of the cars and get warrants to find out who was renting the cars at the time of the crime. Let's also say this was in the days when the records would be paper records in filing cabinets.
Let's say Hertz's files are sorted by year and customer name. To find out who had the Hertz car at the time of the crime someone has to just go through all the customer records for the current year looking for the one that had the car.
Hertz then ends up looking at the records of about half of the people who rented cars from them so far this year in order to find the one person whose identity the warrant asks for.
Let's say Avis's files are sorted by year and car. They just have to go to the current year's records, pull this years file for the car, and then look at the list of customers in that file to see which one had it at the time of the crime.
Avis ends up looking at the records of just the people who have rented that one car so far this year to find the one person whose identity the warrant asks for. That's a much smaller number of people's records than Hertz has to look at.
Those two warrants would be identical as far as the 4th Amendment goes, because what they are asking for, "who had this particular car at this particular time?", is reasonable.
I don't see how it would be any different for records in an electronic database instead of a paper database. The constitutionality of a warrant is not going to turn on whether the query planner can use indexes or has to do a full table scan.
Readit News
[1] https://apnews.com/article/google-search-arson-suspects-colo...
https://help.kagi.com/kagi/faq/faq.html#why-trust
It's weird how easily we've acclimatized to creepy tech-company spying. For Google to snoop on your searches, record them, and permanently archive them for retrieval years in the future was a very creepy and ignominious business model, and it's embarrassing how readily we all went along with it, for the convenience.
It says the same "No activity" for YouTube but I can clearly go on YouTube watch history right now and see every video I've played. That's extremely fishy.
If I go to youtube activity controls and click "turn off" then my history stops being saved (checked by watching a video) then when I click "turn on" then the history starts working again.
https://news.ycombinator.com/item?id=32673075
It would be one thing if they asked for general access to every search that had been done in town. But, wasn't this a case of asking if they searched on a specific address, and it turning up a silly low number of results? And then they were able to find other leads and data?
Or is there a more sinister vibe that I'm just not seeing here?
People are less likely to look up, or say, things in private that can get retroactively interpreted in the worst possible way by state actors who are looking for an easy arrest and/or conviction.
In the past, it was nearly impossible to round up everyone based on what they looked up in an encyclopedia, or on a map, in private. Your pursuit of knowledge wasn't limited by the very real threat of "Can just having curiosity about something land me in prison for terrorism/murder/arson/etc?"
It borders on thoughtcrime and feels like a fishing expedition to me.
> After a house is burned down, they are likely to ask everyone around if they saw or heard anything
To continue this analogy, simply asking is different than getting a warrant/subpoena for everyone in a neighborhood to search through records of their past speech to see if they said anything they deem suspicious.
I can see some concern over this, but just not in this case. If it was, "all people that searched pyromancy" and that included people that were looking up fireworks, for example. Or if it was all people that looked at any house on redfin in that state.
This was a fairly targeted search, all told. "Anyone been asking around about the victim?"
Saying "checking for searches of an address in two weeks" is materially different than "searched for X suspected terrorist organization in four years".
I'm hoping that the town library would stonewall any such request:
https://www.ala.org/advocacy/privacy
That is, there is a good chance something like this request would be honored by any library. "Any chance you saw a person following the victim out of the building?".
When police ask for information in person, the people being asked know about it. When police instead go to Google with a dragnet warrant, the people included in the search never hear about it.
But there's a much more fundamental problem, a problem of constitutional rights. A warrant is based on probable cause to believe that someone commmitted a crime. The Fourth Amendment expects a flow like this: Police get a report about a crime. Police investigate the crime scene. Police go in person asking about suspects. With this preliminary evidence, the police have probable cause to pick suspects and get warrants to search the suspects.
A keyword search warrant is a general warrant. Police get a keyword search warrant. They search people who aren't suspects. Even if Google does the searching and decides which people to recommend to the police, Google does the searches because the government forces Google to search in the first place. The police use the information to get probable cause to pick suspects. This is backward: the searching comes before the probable cause. The Fourth Amendment breaks in this model.
I could see arguments that they need to scope down the time of request for that warrant, as well. But flat out disallowing it feels just as wrong as opening it up for no process.
Deleted Comment
Executing a general warrant compels everyone to “answer the question” in your analogy, which is clearly very different. You keep saying “ask” but here they seize.
Deleted Comment
Supreme Court cases are not about the case. They're about the implications of precedent.
Leaving aside that state supreme court decisions are overturned often, case law tries to be narrowly applied where it can. And this specific case is fairly narrow.
Deleted Comment
Deleted Comment
So the warrant required producing information about a small number of Google users.
> Like geofence warrants, keyword warrants cast a dragnet that require a provider to search its entire reserve of user data—in this case, queries by one billion Google users.
...but they talk about the total number of people's records Google had to look at to find the small number that satisfied the warrant. I'm not sure why they are bringing that up.
As far as I know that's never been relevant when it comes to the 4th amendment. It's the result set that matters, not how many records you had to look at to find that set which can vary greatly depending on just how how you happen to organize your records.
E.g., suppose that two rental cars were involved in a crime, one from Hertz and one from Avis. Police know the license numbers of the cars and get warrants to find out who was renting the cars at the time of the crime. Let's also say this was in the days when the records would be paper records in filing cabinets.
Let's say Hertz's files are sorted by year and customer name. To find out who had the Hertz car at the time of the crime someone has to just go through all the customer records for the current year looking for the one that had the car.
Hertz then ends up looking at the records of about half of the people who rented cars from them so far this year in order to find the one person whose identity the warrant asks for.
Let's say Avis's files are sorted by year and car. They just have to go to the current year's records, pull this years file for the car, and then look at the list of customers in that file to see which one had it at the time of the crime.
Avis ends up looking at the records of just the people who have rented that one car so far this year to find the one person whose identity the warrant asks for. That's a much smaller number of people's records than Hertz has to look at.
Those two warrants would be identical as far as the 4th Amendment goes, because what they are asking for, "who had this particular car at this particular time?", is reasonable.
I don't see how it would be any different for records in an electronic database instead of a paper database. The constitutionality of a warrant is not going to turn on whether the query planner can use indexes or has to do a full table scan.