4
  • Imagine I have a document with confidential information (new technology, stock information, ...)
  • I would like to send it to a couple of people for review
  • They will all sign an NDA
  • But someone leaks it

Suppose I can find the leaked version back, is there a way to sent it in such a way that I could trace and prove who leaked it?

I would imagine having to sent it to a third party who makes a unique version of the document (eg by hiding a key in some images) and who has asked the destination person to provide the key that makes it unique. This way they can certify that that person received that version and that he is the only one. He will also be the only one capable of opening the document if it is also encrypted. I will not be able to reproduce that document, so I can not falsely leak and accuse someone. And I have a third party witness.

Does such a service exist and how does it work?

  • 1
    You can find out who opened your file using canarytoken. For example, you saved your "secret" document using canarytoken in a private cloud. If someone has stolen this file - you will be notified through the mail that a certain person has opened your file with a certain IP. Yes, considering that many people have dynamic IP you hardly know the real IP address of the attacker, but nevertheless you will find out that someone has compromised your data. – stackflow Oct 2 '17 at 16:13
4

It quickly changes from a technology based solution into an armed guards standing in the room solution. You can use a unique watermark on each page, you could make minute changes to characters on a page, but anyone can still retype the document if they are given unrestricted online access to it.

You could approach the problem slightly differently if you make two groups of secret information, the type you need the whole digital document for "this is a photo of bigfoot" and the type you just need to look at and remember "the next wining lotto numbers are going to be 11,16,23,34,43,44". But this only works if someone leaking the first type of information just sounds like a crazy person "The DEA have a photo of bigfoot, I've seen it!".

  • " you could make minute changes to characters on a page, but anyone can still retype the document if they are given unrestricted online access to it." -- This is how SANS detects whether people are sharing their training materials. Even if one sat down and retyped it all, they'd be replicating those minute changes and if you know who had the copy with that specific set of changes, you'd know whose copy was leaked. – Ivan Oct 2 '17 at 16:57
  • @Ivan yup, I was thinking along the lines of removing a few pixels from the character "i" but I think you are talking about making an intentional spelling mistake, or trading out synonyms on each page which can work but you can't do for every type of document. Then you can do one better and tell each person a different thing AKA game of thrones youtube.com/watch?v=hcK1O9MamN8 – daniel Oct 2 '17 at 18:49
  • The spelling mistake would tell who leaked, but you do not have any proof. The sender could also have leaked the document, e.g. in an attempt to compromise someone. So there would need to be a trusted third party that adds the mark to the document and who actually sends it out. Does such a service exist? – EddieM Oct 4 '17 at 8:50
  • @GVDM you then start running into problems with time, if A has a secret, then later tells B (or a third party) then they could have done anything with that secret beforehand anyway, so unless the secret is something that is created with all parties present (like a key ceremony) then you could never know what A did before the meeting. – daniel Oct 4 '17 at 10:30
0

You could use OpenPuff marking feature to do what you said!

Link to OpenPuff software.

Lets say you had a PDF file and you sent it to Alice and Bob. You can use Steganography to mark both versions of the PDF with a 32 character string. So what use does this have you may ask? Well say if Alice leaked the document online but you don't know if it was Alice or Bob you could download the leaked document and check to see what mark you've set on the leaked document! An example of a string used could be. So now if a leak happends you can tell if it was Alice or Bob.

OpenPuff marking

  • 4
    It's worth noting that this has a large caveat attached to it, which daniel mentioned in his answer. This only works so long as the original document you send is distributed (and gets back to you) completely unmodified. Steganography in particular is easily destroyed: converting to another format, retyping the document, printing and scanning, and any other number of processes that can happen in transit for no particular reason can easily destroy the hidden data, making the document untraceable. – Conor Mancone Oct 2 '17 at 14:47
  • Apart from the above there is another caveat. This tells you that Alice leaked but you do not have any proof. The sender could also have leaked the document, e.g. in an attempt to compromise Alice. So there would need to be a trusted third party that adds the mark to the document and who actually sends it to Alice. – EddieM Oct 4 '17 at 8:45

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.