From a civil liberties perspective, government agencies interested in investigating someone can simply buy location data from marketing companies to build a pattern of life or construct a social graph of a target, all without a warrant and without needing to answer anything via FOIA since the data was collected and processed by a non-government data broker. The FBI spends millions of dollars a year buying location info about people from the cellular companies and data brokers... this isn't an edge case.
Sure. No restrictions and I don’t see why there should be — why would some random company in the US be assumed to be any more trustworthy than a company not in the US?
There should be greater restrictions on info release and racking by US companies, but those restrictions should protect against all antagonists (corps, governments) both foreign and domestic. However there seems to be little impetus for protection at the moment outside some minor efforts in california.
Short answer, yes. The concept of "foreign governments" is very US-centric. This tracking is not a US-only concept and the connection of these data points is already being used in non-US markets by companies which cross these boundaries. The only differences with each market is the number of signals available for capture. For example one party operating in the US is capturing DNS activity while in Brazil they are capturing telecom data.
So Vladimir Putin can track all US and EU politicians? Surely they might be interested if someone explained it to them in a way they could understand. Are there any limitations on what countries can acquire the data?
I worked at a isp back in the early 00s and we not only sold logged customer data but we would charge (separately) for the gov to colo a machine and get a span port.
Imagine if someone like Hitler came to power again and decided to exterminate people of certain ethnicity or political interests. He or she could just order or seize a data dump from all those companies and have everyone ready to be rounded up. During the WWII Germans would seize church records and censuses from government buildings to easily find people to arrest and dispose of.
Today he or she would have an app for that...
I think just as likely, governments use this information to try to elicit responses to various stimuli. They treat voters as consumers and consumers as animals that they can train to react via various kinds of repetition and rewards. More granular data allows them to be more precise in these marketing and behavioral manipulation methods.
Off topic, but recently I've been devastated to learn that every single address I (a US resident) ever lived at, my roommates, even my month and year of birth are available on sites like spokeo and fastpeoplesearch. I just don't know where this info is coming from and I want to stop it. Maybe the Equifax breach?
It comes from a lot of places. Voter records is actually a big source, but also I believe some of the various breaches have made it in. And all of those sites continuously crawl each other and trade databases as well.
I had my identity stolen and after a lot of research I determined they were able to gather the info from all of these public sources. I have been trying to scrub myself from them for 3 years now. It's like playing whack-a-mole. Some are quick to delete, only to pop right back up again months later. Some just don't respond. Some require physical mailed requests. A few are openly hostile. You need to submit forms with google to get search results cache updated for each page.
You also need to clear "associates" such as spouse, parents, children because your info appears on their profiles as well. Tedious process!
I believe the best defense is to have a common name or create a ton of spoof info to confuse them.
> It comes from a lot of places. Voter records is actually a big source, but also I believe some of the various breaches have made it in. And all of those sites continuously crawl each other and trade databases as well.
I wouldn't be surprised if a lot of it comes from customer lists (either sold directly by the collecting company for a little extra profit or bought in bankruptcy/liquidation proceedings).
Lots of googling your own name and email address. Finding all the hits and searching for "remove" or "opt out" links (search "sitename+opt out" etc). Repeating the process for close associates, phone number, address.
Then closely monitor. Submit search results cache removal as soon as it changes (https://www.google.com/webmasters/tools/removals). I've found it's best to do one giant wave of removals to avoid propagating from one site to another.
I think many agents are in place since many years tracking everyone, accumulating information leaked from different breaches.
I believe it is widespread today at least in several countries their intelligence really have a kind of "global human database" where everyone is tracked and all digital traces are carefully collected and can be used at appropriate moments for different interventions (like elections, policy change, trade agreements, etc.)
This stuff was always public information. These sites collected it and charged to access it in detail. They basically replaced people who did this professionally. Gumshoe as a service.
You can opt out, but it's a grueling, site-by-site process and they don't make it easy - often you have to do it several times. See https://inteltechniques.com/book7.html for some strategies. It's tortuous and should not be this way - we should really have national privacy legislation, but all the companies who are based on (or just get extra profit from) surveillance capitalism will fight such a thing tooth and nail.
> You can opt out, but it's a grueling, site-by-site process and they don't make it easy - often you have to do it several times. See https://inteltechniques.com/book7.html for some strategies. It's tortuous and should not be this way
I totally agree with you about how unreasonably burdensome the process is (it probably took at least a solid 40 hours of effort for me to get my results mostly scrubbed from publicly accessible sites a few years ago).
However, I think you have to put an asterisk next to "opt-out." In a lot of cases, your data might still be accessible in premium or specialized products (like those for law enforcement). I know of no law besides the California-specific CCPA that actually regulates opting-out, and a lot of these sites like like they're operated by sketchy people.
It's fascinating to me that gyroscope and accelerometer data don't seem to require permissions or explicit user opt-in on smartphones (at least that was true a few years ago and still seems true from a quick Google search, although I'm not sure if things have changed). As a side-channel, I've read about them being used for:
- This.
- Audio recordings of users (https://arxiv.org/pdf/1907.05972.pdf)
- Keystrokes (https://medium.com/@tomasreimers/axolotl-a-keylogger-for-iph..., this one is shameless self-promotion and also why I'm interested in the topic)
To me it seems that there are so few legitimate use-cases for device motion (mostly fitness related and https://www.cyberdefinitions.com/definitions/SMTH.html) that it's almost shocking you can access them by default.
Indeed. I myself have hit significant roadblocks trying to block access to my device's compass. Between storage in a magnetic-clasp case and its car mount not facing directly toward the direction of travel, bad compass data significantly interferes with the ability of my navigation apps to tell the direction my car is going. I wish I could just force them to orient based on my direction of travel.
It is disappointing when articles like this provide only vague/ incomplete implementation details. Is this something we can mitigate? I don't think all apps are given access to this movement data which occurs in the background so it should be something we can disable for specific apps or just not turn on for any apps.
While your app is running, this is true, but if the tracking is limited to times when the app is running, the value is going to be quite limited. They specifically mention getting information about whether you are a runner or riding the subway. If you are out running, you don't have an app open, your phone is in your pocket. I'm pretty sure on the iPhone, getting this kind of data while your app is in the background requires access to the health data which is behind a permissions dialog.
This is why I'm frustrated about the article. It doesn't say how the data is gathered or even really hint at it. The implication is it's a ubiquitous, unavoidable problem, but I don't think that's the case.
I remember reading the privacy policy of Snapchat and refusing to install it because they explicitly mentioned accelerometer data.
I’m pretty sure they (and a lot of other scummy companies) are using this data to track location and/or relationships between users based on correlated accelerometer data.
If you are in a moving vehicle, the accelerometer patterns would be very similar and would reflect the movement/vibrations of the vehicle, which can be matched similarly to how Shazam is able to identify songs based on partial and noisy audio waveforms.
If you have one user whose location and accelerometer pattern is known and then you have another user whose location isn’t known but their accelerometer pattern matches the first user’s one in the same timeframe (with some margin of error to account for clock skew) you can infer that the second user is in the same vehicle (thus location) as the first one. Multiple measurements over time would eliminate any potential errors.
What you describe is totally plausible, and it's insanely frustrating that there are likely smart people who've done this instead of contributing to advancing science, medicine, or at least doing something that doesn't actively hurt society
Chat apps and social media apps are particularly bad at spilling piles of tracking data in general. If you are worried about getting tracked, stay away from social media. iMessage or straight SMS leave some trace, but it's far less than what you leave behind with services like Snapchat.
Obviously people are not going to install an app that is only intended to track them.
However, app makers will put this tracking company's SDK into their app so that they can get a hold of the generated data. But that data is then also shared to the tracking company, and used to build up a complete profile about you.
And, of course, you just know that it'll come bundled in shitty Android ROMs as some darn home screen widget that you can't get rid of, so it'll constantly be "running".
"An app is active when it is receiving events. An active app can be said to have focus. It gains focus after being launched, loses focus when an overlay window pops up or when the device is locked, and gains focus when the device is unlocked."
First, I was a little vague. When I said OS makers, I was referring only to iOS and Android, Mac, Linux and Windows apps are always active when in the background.
As for how they receive notifications while in the background. On iOS (and I'm sure Android is similar), apps register notification events with the OS and the OS sends a notification to the app for that specific event. For example you can register a specific number geographical areas and get notified when a user enters or leaves those areas. Your app is still in the background, but you can perform some small task in response to that event[1].
Your app can also get push notifications over the network or timed notification while in the background. For example a podcast app might get signaled that a new episode is ready and download that episode.
Backgrounded apps can also pop notifications which put alerts up and trigger a sound/ vibration. That's how most notifications get done.
The type of notifications you can get is pretty limited though and the accelerometer is not one of the things you can listen for.
There are also a few specific things apps can do while in the background[2] which generally prevents them from being killed. Media playback is the big one. But again, access to notifications and events like the accelerometer data is quite limited. I believe fitness apps might have access to accelerometer data while in the background, but I'm not sure.
[1] This has been abused and it's been a big game of cat and mouse.
At some point, the experience becomes an endless pile of permissions dialogs. We already see this and see complaints about it.
I think this article is a bit alarmist and overstates the issue. So long as much of this data is limited to foreground apps and background activities are strictly limited, the issue is not nearly as bad as they suggest and as I mentioned previously, I find it unlikely some of the specific abuses they mention are even likely to happen.
No, I think Apple needs to start auditing apps and deciding for users if these permissions are reasonable and banning apps that use sensor data maliciously like this. People choose Apple products because they want apple to make these reasonable privacy choices on their behalf.
At install is really not an ideal experience for most apps, unless the feature is crucial to the functioning of the app (eg. camera access for a camera app).
The right time to ask for permission is when it's used. For example in iOS 14, Signal asks permission to access devices on my local network to sync over Wi-Fi at first launch. I don't want sync over Wi-Fi, so I said no, but many people might just say yes in order to start using the app.
No, because at least on iOS the pop-up has a button for allow and a button for deny and the app must function wether you allow or deny, it can only disable related functionality.
Exactly, which is why Apple should realize a messaging app doesn't need to be sending high-accuracy accelerometer data off device and ban apps that do. Ultimately GDPR laws need to make this illegal. The scary thing is that giants like Google and Facebook are doing this tracking with the kind of resources only perviously available to governments.
> They’d rather not collect the data than send the message that they’re tracking users, “even if they may very well be,” he says.
Because most people don’t have any idea of how pervasive the tracking is and very few people would be ok with it if they did.
I don’t want my phone doing the kind of physical tracking talked about in the article and I don’t know anyone who would, especially for advertising purposes.
Hard pass. I hope their company goes bankrupt some day.
I agree with your sentiment, but tracking is fundamental to mobile devices: for you to receive a call or message, the network needs to find your device. Therefore, your provider knows where you are (within some radius based on there cell tower triangulation anyway). I assume they keep this data and, possibly, monetize it.
> I assume they keep this data and, possibly, monetize it.
I assume you're correct, and find it unfortunate that we don't legally prohibit this practice. They really shouldn't be allowed to use that data for anything other than providing the specific services of connecting calls, providing internet access, etc. But it doesn't seem like enough people care to make this a reality.
Even if your provider knows where you are (and Librem 5 has a hardware modem kill switch for that), it does not mean that the software can send your location to everyone, too.
I am genuinely interested in this, so I would appreciate your response. My assumption in the comment was that the phone provider keeps a log of device location information. It needs that information to cause your phone to ring. This is completely independent of any software on the device. Am I missing something?
You are right, this is what happens. However, with Librem 5, you can switch your modem off at any moment if you need privacy. Obviously, calling won't work when it's off, but you at least can balance your privacy with the availability for calls.
At the same time, the phrase "tracking is fundamental to mobile devices" does not tell the full story. Yes, it is fundamental for calling [0], but the parent comment said:
"most people don’t have any idea of how pervasive the tracking is and very few people would be ok with it if they did".
This is not (just) about the modem tracking, but also about the tracking built into the OS itself. There is no such thing in Librem 5: https://www.youtube.com/watch?v=yHcHi0TBFv4.
[0] and tracking is not fundamental for IP-native communications, which is what Librem 5 is trying to make the default option.
Upd: "the phone provider keeps a log of device location information. It needs that information to cause your phone to ring"
This is also not true. The log is not necessary for the phone to work, only its current location.
The phone provider knows, within a certain margin of error.
I would expect (not welcome, but expect) that that data is available to law enforcement. Not random commercial companies who want to sell me stuff.
What's described in the article is creepy as fuck, using spyware to literally spy on everywhere we go and everything we do, for financial gain, for sale to anyone.
With 5G, I believe that this margin of error will be significantly smaller. But either way, with wifi and gps your location information is pretty accurate already.
uBlock Origin blocked 34 requests/elements on that page alone. Requests to Amazon, Google, Oracle. Tracking pixels, segment tracking, Ad displays, cookie notifications, you name it. This is all on a computer web browser where data collection is somewhat limited when compared to a mobile app.
After scrolling to the bottom and back up, Brave blocked 69 (nice) “ads and other creepy things”, on the phone. I honestly barely go to sites like this anymore.
Big media is just gross. The quality is dirt-poor, the business models are abusive, and as the Facebook era draws to a close and the money dries up, the production value is trash too. Ars is good in this category, but still bleh.
If this was 30 years ago, Ars would have been a print magazine you'd have to pay maybe $4.99 for. And they'd still struggle.
Magazines with ads and subscription-based models went out of business all the time. The ones that survived either appealed to the widest possible demographic, or had a deep-pocketed benefactor keeping them alive.
I disagree that the quality is dirt-poor. I can still find a decent amount of good articles on Ars Technica, but they also have a lot of noise dedicated to generating clicks and likes. The article posted above is a good example.
In most cases this comes from mobile ad/analytics frameworks (Facebook Ads/SDK, Google AdMob, StartApp, Chartboost, MoPub, etc. etc.).
They usually ping back with usage counters and probably other analytics data. In many cases those SDKs even join data between apps on your device (e.g. you give your age to app A, your name to app B, your location to app C and they merge it on the backend). This data is then shown to users and in a lot of cases also sold in bulk.
Note that most of these frameworks target iOS first and Android second - due to population and revenue.
I ran a similar app and it was surprising how many requests were getting made, it wasnt super helpful as they were all either AWS or gcloud requests. I did see which apps were making the requests and at what frequency, but that was about the level of detail available.
This app requires a rooted phone. I don't have root on mine. Is there something similar that doesn't require root?
My Pi-Hole is useful for tracking/blocking via DNS, but I can't tell which apps make which DNS queries. I'd imagine you must be running on the phone to get these data, or use deductive reasoning/matching/etc.
This one's obvious. I've gotten ads ON OTHER PEOPLE'S DEVICES targeting me (a habitual blocker of ads). They were very industry specific and I was hanging out with people that would have nothing to do with that industry.
> "The old world of these predefined segments like soccer moms or other [ad] categories will start to decrease"
Every time I hear someone in the ad industry talk, I have to laugh. "Soccer moms" is a stereotype, not some scientifically defined demographic. And they're targeting based on this after years of tracking and data harvesting?
When media organizations call up their contacts for a quote, more often than not, they'll get a stock quote that could have sounded somewhat relevant 10 years ago. People in the know about industry's state of the art are not in the habit of just giving that information away to the papers.
What does "fighting" mean in this context? Money will not change anything if the lobbyists on the side of the advertising behemoths continue to have the ear of lawmakers.
Facebook and Alphabet are two of the world's largest corporations; they, and the giant cottage industry of analytics and ad arbitrage companies are entirely dependent on maintaining this status quo.
Even Apple, a company not reliant on ads, succumbed to advertiser pressure and held off on introducing changes that would limit their ability to collect tracking data.
> Google and Apple have taken steps this year they say will help users shield themselves from hundreds of companies that compile profiles based on online behavior
In other words, Google and Apple are using their platforms to eliminate advertising-profile competition.
Good point in the article that barring apps from aggressively tracking users (which is a good thing IMO), creates more power for companies like Facebook/Apple/Google/Amazon that already have access to the data
Even if this is true, I would suggest that you just try to find another job in the meantime. Who knows, maybe you can bring a good change to the world, however small, by your action.
Fun fact: Apple tracks your location and keeps its history, too: https://support.apple.com/en-us/HT207056, see "Significant locations". They claim that this info is only stored on the device. My watch is always in Airplane mode, bluetooth is always off, only wi-fi is on. And the watch actually has a pretty precise location history in the settings. How?
The whole goal of Google/Apple "increasing privacy" in this way is to basically have a monopoly on user data, and thus gate any mobile advertising behind them. This is all revenue based, nothing to do with actual privacy.
When I can take a smartphone, uninstall any apps that phone home, and then block/uninstall any apps that listen for external traffic, all without breaking core os functionality, then I would believe that those companies are truly concerned with privacy.
> When I can take a smartphone, uninstall any apps that phone home, and then block/uninstall any apps that listen for external traffic, all without breaking core os functionality, then I would believe that those companies are truly concerned with privacy.
Which Apple watch? "Apple Watch Series 2 and later and Apple Watch SE have built-in GPS that allows you to get more accurate distance and speed information during an outdoor workout without your paired iPhone."
Airplane mode does not turn off GPS. So it seems very reasonable that an Apple Watch would be able to know where you are. This doesn't prove or disprove what happens with that information, but it's a perfectly reasonable way for your watch to know where you are.
This sounds logical, but whenever I switch Wi-Fi off, the tracking switches off too (no new "significant locations" added anymore). In addition, the list of locations contains addresses, not GPS coordinates, so the connection to a server is definitely necessary even if this is GPS. My point is that the data is transferred to the Apple servers.
No, I don't think I ever agreed on that. It's buried very deeply in the settings (Settings -> Privacy -> Location Services -> System Services -> Significant Locations). You however can opt out.
It's not enabled on any of my iOS devices, so that's one data point. They were upgraded to iOS 13, if that makes a difference (maybe it's opt-out on a fresh install?). I tried searching online and I found an article saying it's opt-in[1], but also found anecdotes similar to yours about it being enabled without their knowledge.
I think it's a little ironic people are so concerned about this kind of thing because so much of what annoys us about advertising is ads that are not relevant to us - due to the limits of targeting.
Ads that are extremely relevant are not even considered adverting to us in the moment. Wont better targeting provide more of this sense and a perception of fewer ads overall?
I don't want to be sold products. I don't care how "relevant" it is to the particular subset of people I most resemble. I want precisely the information I asked for, and no more. Every bit of information extraneous to that, is another piece of garbage I have to wade through to figure out whatever I actually wanted to know or do. The most frustrating ads are ones that don't make themselves clear that they're trying to sell me something until the end and waste my time, when I have zero intention or desire of ever buying something to solve a particular problem. Or the websites that write content that is almost useful, just to get as many eyeballs on its ads as possible. Ads are the slurry of microplastic on the sea of information, of benefit only to whoever dumped them in the first place.
And if I'm enjoying a fictional work, I certainly don't want to be interrupted with another person hocking another piece of junk, or worse, integrating the ads into the story and having Gandalf trying to sell me on Pepsi.
> I want precisely the information I asked for, and no more.
Are you willing to pay for access to that information?
Because ultimately whoever is providing content wants to be compensated for their time.
I agree with your sentiment, but there is also a pragmatism here. (FWIW I am willing to pay for, ad free content and don't believe tracking/ profiling should be part of advertising or even legal without consent.
I'm essentially in the same boat as you -- I'm willing to pay for content as I do for my books, or support free sources of information as I might for Wikipedia. The siren song of the advertiser is hard to fight though, even if many people are willing to pay the price for a quality product. The Faustian bargain that I presume leads to this mess: Any ad-free informational product could probably reach a wider audience and thus pump up revenue, if they traded some of their up front cost in exchange for a few modest ad slots or selling a little customer information here or there (To some degree paid users would put up with these things even without any subsidy to their subscription whatsoever, so it could also be directly pocketed).
If I complain about poorly targeted ads, it's because I am shocked that the targeting can be so useless even with the disgusting level of surveillance.
It is not an invitation for even more intrusive spying. It is an invitation to give up and leave me alone.
Ads and advertisers insinuate themselves into every goddamn situation, they destroy trust in human communications by introducing a profit motive f*cking everywhere, and they degrade every experience they're part of.
Was searching for accuracy in modern sleep trackers and smartwatches. Couldn’t find anything except a youtuber from Vienna which had an amazing confusion matrix to show in a video. All the articles I found where basically:”buy buy buy! Look at the features! Oh heres the affiliate link...”
I remember having a Polar heart monitor that said that my table had 67 beats per minute... best of class, so many features!
Until a company starts using this same data to screen prospective employees for unionization risk, or uses purchasing history to identify and filter vegans from slaughterhouse jobs, so there's less risk of those animal cruelty videos getting out. Or show you higher prices online if they determine you're willing to spend more [1]. Or...
Or are they doing that already? How would you know?
Is that a new thing? Seem to me that's been happening for like hundreds of years. I've worked at a PR firm and in a Newsroom. I've seen the sausage being made.
The key phrase was "micro-targeting" not "propaganda". Previously, if you put out an ad full of religious zealotry to appeal to fundamentalist Christians, you would risk turning off more moderate parts of your base. Now you can make sure only the "right" people see that message.
That's because Facebook sells ad space. There are companies that don't do this, and are happy to sell your information. Like, say, the California DMV...
Sure, CC companies have been doing it since the 1950s. But only aggregate data has any value. Advertisers care a lot more about finding a large group of potential buyers than individuals.
Something like the more data you have the less detail you know, because noise expands faster than signal.
If you're implying that advertisers only want aggregate data to build a marketing strategy that requires leveraging the individualized data from ad platform holders like facebook, that feels like some serious cognizant dissonance about the meaning of privacy.
> Wont better targeting provide more of this sense and a perception of fewer ads overall?
No. If you're say, a video game lover, you're still a potential customer for a number of things you don't really care about. Ads don't target things you like, they target things you might be convinced to buy. If you want to find interesting, relevant ads, you'd be better off having no targeting, and finding domain relevant ads. E.g. video game ads on video game sites, rather than ads for the toaster you bought on amazon 3 months ago or women's underwear because your spouse made a quick search on your phone.
I like plants. But I don't get ads for plants, instead I get ads for stupid plant-related gizmos. Why? Because the plant "market" is not online, and is much less lucrative than the gizmo market. There's no amount of tracking that can fix mismatches between commercial and human interests. Or to put it another way: the best things in life are free, ad space is not.
What I'm describing is the best targeting that a massive dedication of the top engineering talent in the planet is capable of. Perhaps if we gave them even more personal information, like a list of all my purchases, they would know not to sell me the toaster I already bought. That doesn't mean there's a market for "good" ads that could reach me instead.
Previously I used video games as an example. Let's choose something more mundane. If I have a pet cat, I still don't want to read ads about cat food. I don't want ads about litter. I don't want ads about stupid cat toys. Or pet cameras. Or pet photos. Or pet grooming services. Or more cats. I don't want your fucking ads at all. Your targeting isn't wrong, but that doesn't make it pleasant. Everyone needs toilet paper. I don't want toilet paper ads.
And then on top of all that, cat product companies get outbid by some sketchy financial company trying to get me to refinance a mortgage because the margin on that is way higher.
I bought a software product for work several times. But it was not purchased in any way related to my personal Google account. Using Youtube to learn about the software has cause me to see thousands of ads for the software I had purchased several times already.
This is poor targeting (especially since it's direct response), and a lot of people I talk to get very annoyed telling me their experiences like this.
That's great for you, but is a tiny fraction of the problem being outlined. If you could say "Don't show me ads for this again", and total ads were held constant, you would not suddenly be inundated with ads you enjoy seeing. Ads are unpleasant, period. Targeted ads are no better.
I find targeted ads extremely creepy. It's like I'm being stalked. I now turn off personalized ads wherever I can.
Any politician that talks about regulating what data can be collected and how it can be used will probably get my vote. I want to see my browsing data protected as well as my video rental data.
> so much of what annoys us about advertising is ads that are not relevant to us
What ? How do you even decide what is relevant for you and what is not yourself ?
I never think about ads being relevant or not, they are always annoying. My attention is not unlimited and I certainly don't like others profiting from it.
edit: Well I guess we could agree that if I get ads and I'm in category "I want 0 ads" it is indeed a case of poor targeting.
