Bitcoin Forum

Can I put a hash of a photo into the blockchain and use it later in court to prove that particular photo was taken before time x?

You can put the hash in the blockchain and prove it was taken before that block was generated.  It would be good proof to me, but proving the timestamp of that block and what your hash proves would likely be "interesting" in court.  That's a lot of math you'd have to explain to a jury.

Courts have "experts".  You just need one that understands bitcoin and can testify that he knows that including the hash retroactively isn't possible.  The other party's expert won't be able to claim otherwise without perjuring himself, and any math teacher or computer science professor from a local college will be able to back your guy up on the tricky parts (hashing), even if he doesn't know bitcoin specifically.

Oh, but keep in mind that the timestamps aren't exact.  There is a several hour window on either side.  That is enough for most things, but not enough for precision timing.

You can prove that particular image existed before the block was added to the blockchain. However, there is no way to prove the image was not manipulated with tools like Photoshop.

I didn't wait to see if it would be accepted in court: https://bitcointalk.org/index.php?topic=39548.msg519255#msg519255 (https://bitcointalk.org/index.php?topic=39548.msg519255#msg519255)  ;D

@OP: expert witness is all you need

@Frozenlock: Love that link. Sending yourself sealed "original work" like screenplay before going to Hollywood to show it to local copyright whores is so 20st century.

commitcoin (http://people.scs.carleton.ca/~clark/projects/commitcoin/)

ha! actually, that was exactly what I had in mind. I'm currently moving into an office and a window is damaged. I wanna be able to prove it happened before my time there ;)

You can present it as evidence with an expert witness, whether a jury will get it is another story.

I'm sure court investigators have had many chances to hone their photoshop detection skills these past two decades.

is this implemented anywhere? the paper blows my brains out just by looking at it.

"hash" is the 2nd most important 4 letter word you need to understand, kid.

http://en.wikipedia.org/wiki/Cryptographic_hash_function

The point is it could have been manipulated before hashing.

For the purpose of documenting the fact that damage occurred before a certain time this technique works just fine.  The other side would have to argue that you photoshopped up some fake damage intending to do the actual damage later.

you can use it in the way described in proof of concept, but it burns money. I will implement it in a month or two as part of my thesis, however.

It's evidence of Transactions occurring between addresses. Period.

To explain the intent or purpose of those transactions is a different story.

Can it be done, sure. But if LE is at that stage in investigating, they don't need the blockchain. It will just be another aspect of proof.

Think of how hard that is. Even if you prove that TX1 went to TX2, WHO sent TX1 and WHO received it and for what purpose. And now this gets even more complicated because the route from TX1 to TX2 may have thousands of transmissions between them.

Then you have the encrypted transfer of wallet.dat files containing an amount of BTC which would bypass miners and fees. I.E. There will be 21 Million BTC. Now imagine 21 Million Wallet.dat files being transferred to others bypassing miners all together. And the method of transfer doesn't even need to be electronic.

Not to mention that if it becomes accepted, people could send BTC to someone to implicate them in something that they had nothing to do with. Or cause them headaches.

For Example: PP customers send PP's CEO a $1 with a note: Thanks for the BitCoins   The PP CEO would be wrongly assumed to have violated his own companies policies. (But that wouldn't be fair or correct)

However there is a good point here:

The qt client allows the viewing of Transmissions and Receipts whether the wallet.dat is encrypted or not. This, IMO, is a privacy and security flaw. Anyone with access to the program could view your personal financial transactions.

To me this is akin to allowing everyone to see your online banking account details but they aren't allowed to transfer anything. If and until this is fixed, I would suggest for any serious and private TX to use the program from an encrypted disk and/or folder.

I'm sure I'm wrong about something, correct if necessary.

Dude, good job on that. I will include this in my presentation at Future of Money Summit if you dont mind.

commitcoin isn't my idea, to clarify. I'll just do the implementation, since there isn't one yet, as far as I can tell.

Heh, thats just as good. If you can have a working beta within 4 weeks, I'd love to use it in the presentation  ;D

@BTC_bear very interesting questions, thought provoking. I'd start with a claim that private keys are secrets and that TX1 and TX2 could only be initiated by a person(s) with knowledge of the keys to sign ... and here are the keys.

I'll do my best. It's actually pretty straightforward to implement (especially using something like bitcoinj), but there's so much (non-bitcoin related) stuff to be done alongside that... :)

Thanx.

 Just pointing out that private keys in no way ties it to a single entity or person even. A person's computer might have been compromised, or even by intent more than one person might have a copy of the private keys. I know of one business entity that does this. In fact, the QT Client does this.

Supposedly there are 3 persons with the ability to send a message to the client. (Gavin, Theymos, and Satoshi) Who sent the recent message to the clients about the security update?
 

Like what, design?

no no, a few other totally unrelated projects.

I posted private keys in semi-public forum to challenge users to grab a bitcoin and then refilled the address to see how fast they notice and grab it again. I have even no idea how many wallets have now that key. but that was intended and I would not use that key to exercise any kind of serious TX (like writing to blockchain for future use in court should be)

the above example of a collective knowledge (especially the keys from bitcoin message function), I assume the key has been changed so Satoshi no longer knows it, unless Gavin and Theymos are married to each other, they should be able to testify against each other. If the key is the same since project inception, then it might be a problem if both deny writing that message and Satoshi would not show up in court.

I have multiple pgp keys and I share some of them willingly and proactively (cyber last will, trusted friends hold copies to settle my affairs in case of brain damage or death) but I still consider them as personal and do all I can to keep them safe and secret. some of the keys have even my name in them : )

Then I think we agree that it is a bad idea to try to use it in court. As one can foresee the Prosecution telling the Jury only the defendant could send the money/msg. Then 'someone' else uses a copy of the wallet/private key to send money/msg and destroy the prosecution's case.

I'm willing to further discuss and finally probably also to agree that an indisputable link between a person and a transaction can not be established (issues with copies, backups, ...).
But (and I quote OP here) I would continue to argue that a hash of a digital file (a photo in this case) can be found in the blockchain and also that it can be proven when it has been included in the blockchain.

Why it's there, who's benefit it is I don't discuss. Just the fact that it's there and when it was included. Would the link between the file and it's fingerprint in blockchain stand?

edit: but my initial argument in this tread, you disproved it

Exactly right.

Unless one or more of the major functions used by bitcoin is seriously broken, then the existence of the hash in the block chain would easily and convincingly prove that the document existed prior to the timestamp on the block (more or less) and that it has not been modified since then.

It could still have been modified prior to being hashed, and it could have existed for a long time prior to being included, and anyone in the world could have posted it.  But it existed, in that form, at that time, and you can prove it.

Forgoing the former as that is back to intent and purpose, the latter can be done within a mathematical certainty if the chain is verified as being correct. (The actual current chain in use and not an alt or forked chain) Even without timestamps, the blocks occur approximately every 10 min on average so one could give an approximate time frame with only the time of the first block being confirmed.


btw: Files have been encoded into the blockchain. But with time, no encoding would be necessary as the data will be there. One could just write a program to pick out the data necessary to create a file.  Simple example: Rather than encode 'Hello World', just find an ascii equivalent in the chain and point the reader to those addresses to decode. Hence, someone else unintentionally encoded the message, the reader just picks the right address to produce the desired message/file. Granted this is easier for simple messages and very difficult for harder messages or images but possible.

IMO

and offer some publicly accessible service with it?

mila, btc_bear: you are discussing "blockchain in court" in more general terms than the OP asked for, correct?

because using it to prove some data has existed before time x does not involve ownership of keys and such, since it doesn't matter who inserted the hash, right?

https://bitcointalk.org/index.php?topic=74863.msg829561#msg829561
well, ^^ here I was returning from the general topic detour if that link between a file and its timestamp would stand in court. got confused by the reply (forgo the proof unless answering qui bono) and thought about the seti@blockchain for a while but was unable to contribute to the talk.
I mean finding a hash of anything useful in the blockchain should be harder than finding a blockchain (unless I have the file, know which function to use to get its hash and the method how it's encoded in the blockchain).

so refocusing back to OP, I think it is a proof. while a use of a commercial authority to provide something like XAdES-T (signature with certified timestamp) could be easier to justify in a court, blockchain would do basically the same without revealing who made the request in the first place

you can simply register a name of up to 253 characters with namecoin. timestamping - piece of cake.

website forms for submission/verification, most likely.