6

I would like to be able to prove that I wrote the article that resides at a given URL on someone else's blog. Is there a way that I can do this with public key cryptography? I know a bit about the topic and I am thinking that I could:

  • use my private key to generate a public key which I will display on my own website for reference
  • use my private key, seeded with the URL of the blog, to create a new public key for each new blog URL I write on
  • ask the reader to verify the authenticity of my work by somehow comparing the two public keys

Am I on the right track? It's the somehow I need help with.

3 Answers 3

6

What you can do is use PGP to sign your text (another article here). You will then need to publish your public key so people can verify that you have "written" this text. Note that there is no way they can actually verify that you have written the text (you, yourself could also just copy the text from somewhere and sign it, even though you have not written it).

  • make your public key available (preferably you give the key in person to everyone, in reality make it available on a page solely under your control)
  • sign your posts with your private PGP key and make it available in the blog post or on a personal website
  • If people want to verify that you have written this blogpost they take your public key and verify the post. If they succeed they know the post was signed with your private key. This means that the only person who could have encrypted the message, is the person who has the private key. Considering it is your private key (you never give this to anyone), they know it must have been you.
10
  • that sounds like what i am after. how would people verify this? ie what is the process of comparing the public key and the signed post? Aug 26, 2013 at 9:17
  • Added the process Aug 26, 2013 at 9:22
  • excellent. i wasn't sure if the post could be decrypted with the public key alone. thanks! Aug 26, 2013 at 9:31
  • 1
    Note that it is not decryption, but signature verification. Aug 26, 2013 at 11:13
  • 1
    It actually doesn't prove you wrote it, rather it says something about your approval of that post (provided the private key isn't stolen, of cource :)).
    – peterph
    Oct 1, 2013 at 9:44
7

After the fact (and limiting the answer to viable technical solutions), there is probably no way to prove you wrote something. I will assume the content is yet to be published.

There are two parts to a reasonable "proof", the first being association of the content with a digital identity, i.e. signing the content so that it can be verified with a recognisable and easily available public key. The second (which may not be required, depending on how paranoid you are) is proving you created it first (i.e. to prove or disprove plagiarism). A digital signature provides integrity and authenticity (proof of sender). Proof of authorship (related to non-repudiation) is a more difficult problem to solve.

The simplest way then is with a public key solution, as you suggest: before you publish you get a trusted third party who provides a secure timestamping service to timestamp and sign a digest of your signed content. The timestamping service will use X.509 certs, you are free to use PGP or X.509 to sign your content here.

Your "provable" published data should be: the content (signed by you) and signature (if not inline), your public key or its location, the digest method, and the third party's timestamping signature (of the digest).

This question on signing code commits pretty much covers the technical details: https://stackoverflow.com/questions/11913228/how-can-i-use-rfc3161-trusted-timestamps-to-prove-the-age-of-commits-in-my-git

A possible problem here is how to represent your content in a canonical format so that others can verify the signature. You should consider signing the content in a format which can be downloaded "intact" and not subject to unintended reformatting or alteration by the server side blog software. Offering a PDF version of the article would be a solution, and you also have the option of using an X.509 signature and timestamp within the PDF itself (PDF).

If you're old-school, there is a free 21st century equivalent to posting yourself a copy by registered mail, which may or may not be useful, and probably has about as much legal weight (i.e. negligible): http://www.itconsult.co.uk/stamper.htm


There is now a similar question on Programmers.SE , one of the interesting answers involves using Bitcoin to make a sacrificial micro-payment to a non-existent address (your content hash value). Once your transaction gets onto the blockchain you have your "proof" (albeit with reduced timestamp resolution, it could take some tens of minutes).

2
  • It's possible to study the author writing style and verify, with a certain level of confidence, that the writing style of one author is the same that is found in some published article... Aug 26, 2013 at 13:29
  • @woliveirajr, But that certain level of confidence by itself is semi-useless if it's a low number.
    – Pacerier
    Jun 4, 2014 at 17:26
3

I think along with the who, the when is also important. This bring into play timestamps. These are a third party service and they sign a hash of the text and add a timestamp. This proves that the text was in existence before that instant of time. So even if there are copies, your copy was the first one (or not).

unlike gpg, timestamping services may not be free.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .