Hacker News new | past | comments | ask | show | jobs | submit login
Deanonymizing Darknet Data (atechdad.com)
37 points by julianj on July 17, 2015 | hide | past | favorite | 12 comments



The data is not anonymous to begin with if it contains plain identifying information (like an email address).


> Parsed hundreds of thousands

> found 37

I agree that anonymity is hard and when you make waves it becomes harder but it would be tough to draw that conclusion from those stats. Interesting nontheless.


There would be many more than that if markets didn't (usually) strip metadata by default.


quite possibly true. I can't be asked to dig it up but either this post, or another person speaking about your data said these sites are hand coded and often don't scrub data from uploaded images.

if you as an individual can't be asked to scrub a bit of metadata off of a photo you upload to an online black market you probably aren't qualified to work part time at walmart. dodgy bit here though, it would be pretty trivial to edit it and intentionally leak false info and be moderately clever while everyone thinks you are daft.


I wouldn't say often. Most of them, as far as I can tell, get that right from the start, and the ones that don't fix it fairly quickly after being told.


Incidentally, 37 is definitely low. When I did the same thing in early 2014, I got >90 unique GPS coordinates.


Why the discrepancy? Just that most images on the sites aren't jpgs? (What else supports exif, tiff?)


I believe I checked PNGs as well, and I went through all my archives up to that point (it wasn't just Evolution which had forgotten for a while to strip image metadata). Possibly my grep was also better at checking variant field names? Dunno.


Probably a combination of that, more awareness of this issue and more hosts stripping exif data from uploaded images.


Wow. What an accomplishment. :/


TIL stop allowing your pictures to have meta tags.


Most online child pornography is not hidden nor encrypted in any way. Just search Bing; you can get the keywords from wikipedia then follow bing's suggested queries.

Microsoft claims that they remove child abuse links from their index upon notification by such organizations as the National Center For Missing And Exploited Children but clearly they dont.

There are some imagehosts that either have no way to locate their images from their homepages, or whose search forms dont list child pornography in their results, however its all indexed by bing.

The videos at the filesharing services commonly have obvious or easily brute-forced passwords. There arent that many different passwords in widespread use; I expect a list of the 100 most common passwords would decrypt 99% of the videos.

The passwords and lightly obfuscated links are mostly posted to dead forums. If you operate a forum yourself the simplest way to defeat this is to disable posts in threads older than a month or so.

If you want to find these posts on your server look for a large number of occurrences of the word the word "password".

I have even seen this at last.fm.




Join us for AI Startup School this June 16-17 in San Francisco!

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: