4 ms·
Why is this surprising to anyone? It is clear that ads use tracking mechanisms and cookies and this is no different. Audio feature detection isn't even a novel
by inglor 7y ago
Why is this surprising to anyone? It is clear that ads use tracking mechanisms and cookies and this is no different.
Audio feature detection isn't even a novel techique.
I've seen trackers look at download stream patterns to detect whether or not BBR congestion control is used, I have seen mouse latency based on the difference between mouse ups and downs in double clocks and I have seen speed-of-interaction checks in mouse movements.
Just checking for the constructor of something an ad might legitimately use (like audio) is relatively benign to be honest and it is naive to expect ads to not do this and it is why I use an ad blocker even on sites without annoying ads
- inglor 7y agoAnd as a fun fact networking timing fingerprinting attacks and work even if you don't have JavaScript enabled and I have been able to make a PoC that was very accurate (I did not release it but I did disclose some bits to relevant parties)
- saagarjha 7y agoI hope "relevant parties" includes "browser vendors" and not "adtech companies" :)
- inglor 7y agoYes of course, browser and OS vendors.
- function_seven 7y agoWhich one is Google?
- ehsankia 7y agoBut for code that's supposed to be so smart in trying to fingerprint people without them knowing, calling an API that throws a warning in the browser seems like a really stupid move. Especially since that can be checked through feature detection, which is literally what this code is doing...
- teraflop 7y agoOne reason it's surprising is that, until recently, SO was particularly resistant to allowing invasive and/or obnoxious ads. See also the recent decision to allow animated banner ads on various Stack Exchange network sites.