Provenance and Copyright

Ming Thein has a hilarious little sequence of posts which are mainly about hinting at conspiracy against him, and talking about his high level contacts at some watch maker. Long story short, someone made a mistake, flickr pulled some of his photos, mistake corrected, photos back in a timely fashion. No. Big. Deal.

But it does raise the question of provenance in this digital world. I think most of us are at least vaguely aware that proving ownership of a photograph (in the legalistic, copyright, sense) is pretty hard. There are some vague suggestions about "original RAW files" and Ming illustrates a hilarious faith in "uneditable file formats".

I can edit any file. Any. File.

I cannot forge cryptographic material. In general, I can probably find an end-run around any complex crypto solution for proving who took a picture, when, though. Consider this, which is both a real thing, and an analogy.

One offered solution is to register your photos with the Copyright Office, which I guess is a real thing. Then to prove that you hold the copyright, you simply check with the Copyright Office, right? That sounds foolproof to me, some trusted third party. Except, as far as I can tell, there's no barrier to me registering your photos with the Copyright Office. The penalties are, one assumes, horrific, for this illegal act. So what? I can still do it, and if either I live in Kazooikstan, or 1,000,000 people a day are doing it, the penalties are irrelevant. Similar approaches will usually work with complex cryptographic software solutions, because most people who design these things simply aren't smart enough to get it right. It's fiendishly hard to get it right.

So how do you prove you took a picture, and not that other guy?

Well, if you had a negative, that would be something. One can simply shoot a copy of a print, and get that, though. Perhaps if you had a proper original negative, that actually has more detail in it than is realized in the print? That has original grain in it, that one can see in prints but which is verifiably original grain? That would be extremely hard to fake, and would probably constitute sufficient proof for even the most finicky court.

If you had a RAW file that also had detail not visible in any JPEGs or prints, that would also work.

Of course, if I can get access to any digital format of your picture that has all the information in it, I can construct a RAW file, and your goose is cooked.

So you need to hold something private, hold close something with visual information nobody else has. Then, to prove it, you need an impartial third party to verify that the visual information in your private copy is consistent with the rest of the visual information, and therefore probably not forged. You need an expert to verify your claim.

This isn't scalable, at all.

If there was any money in it at all, there would be a brisk cottage industry in constructing forged RAW files, in registering other people's photos under someone else's copyright, and so on. The courts would almost certainly punt, and do something akin to this: we're only looking at cases with registered copyrights, and we're deciding for whomever registers first, period.

This isn't happening, thankfully. It tells us something, though, about the true value of copyright. If nobody wants to steal it, how much is it really worth? People steal photos all the time, but I've literally never heard of anyone stealing, or even making a credible attempt steal, a copyright.