They don't have to be running. Here's my understanding (for iOS anyway):
1. Apps have the ability to subscribe to bluetooth callbacks from the OS, which is constantly scanning for them (about once a second, from memory). It will be something like, "wake me up when you detect that beacon with UUID ABCD123 is in range". ABCD123 would be the standard ID of a marketing company's beacons - there could be millions of them.
2. The beacon also have sub-IDs identifying the exact beacon being used. The marketing company will know which are where.
3. Whenever the beacon is in range, the OS pings the app with the data, which decides what to do with it in the same manner as a background data refresh. This could be something useful, like waking up to let you know your suitcase is nearby - but it could also be silently uploading that data to a server.
4. The software to do this is being bundled as a paid SDK in a great many seemingly-unrelated apps, such as weather apps.
5. This behaviour is not counted as location services in the OS, and may or may not be disabled even when bluetooth is "off" on the phone
Corrections welcome but I believe that's roughly what's going on.
I'd greatly appreciate something like Little Snitch on the iPhone so I could see which apps are doing this and delete them with extreme prejudice. Back in reality, I'm glad this is getting attention - at the very least Apple should be providing a list of apps requesting BT access, and indeed any network access over time.
I'm pretty shocked all this can go on in the background without permissions. With Apple adopting such a privacy-conscious stance, I really hope this gets their attention so beacon scanning requires explicit permission in the future (and separate from location services -- my weather app needs to know where I am, but certainly doesn't need to scan for beacons).
This is honestly pretty egregious.
It certainly is. I knew about the mechanics of it but hadn't realised it was being so widely abused. I would like to see Apple come down on this swift and hard.
Third party paid SDKs! Those cunning bastards. AdTech really is the dregs.
It would be trivial to do this if you can get your hands on an iphone with a jailbreakable OS version.
1. Apps can be running in the background. I'd say its more common than you think 2. You need to provide location permissions (either the ACCESS_COARSE_LOCATION or ACCESS_FINE_LOCATION). It seems like every app requests these permissions anyway, so not a red flag just on its own. 3. Something scanning for a WIFI devices has no way (easily) to coorelate those addresses back to a user. With this bluetooth low energy method, the users own phone is the one who reports the detection back to {company}'s servers, so it can pass along any/all other information it has about you. Their location accuracy is also pretty crazy (radius is within centimeters if they've done it well). 4. Bluetooth low energy is actually crazy easy to set up, see here for more... https://developer.android.com/guide/topics/connectivity/blue...
1) Some random third-party app has to be running on your phone to detect beacons and send the data back... how viable/likely is this actually? It seems like this would only ever effectively detect a tiny percentage of users at best who just happen to have one of the apps open while walking around a store?
2) For an app to detect beacons, don't you have to give permission for the app to use Location Services? I've tried Googling it but can't seem to find a definite answer... I'd be surprised (and saddened) if Apple or Google are allowing apps to detect beacons without explicit location or Bluetooth permissions.
3) If the goal is to track as many users as possible... wouldn't it be far more efficient to look for Wi-Fi devices that are scanning, and identify them by their MAC address? I don't understand what Bluetooth beacons enable that Wi-Fi scanning doesn't.
4) The article lists companies that provide these third-party toolkits... but not a single name of an app that uses them, or what percentage of phones contain an app with them. Since this is the main accusation of the article... I don't understand why they wouldn't provide even a single instance of proof.
I've just seen a lot of very questionable reporting from the NYT in the past on tech/security/privacy, so I'd like to understand better how real this is or not.