In the (old?) days of film, if there was a controversy regarding paternity of a picture, the slide or negative would easily solve the matter. If you managed to get the infringer on trial, you won. A duplicated slide or a reshoot from print is easily detectable. Nowadays it would be even easier because you give scans to the client and keep the slide always with you (a reason not to switch to digital).
In the modern days of digital photography you have no negative to show the judge. You have the RAW file but that is just an identically duplicable file like any other computer file, and we have to consider that many professionals give their clients the RAW files.
Another thread on another forum (a photographer had accepted to do a job for free and without contract for a magazine, at a friend's studio, but wanted proper mention on the magazine. The photo was published double page, no mention, and no contract either, so the poor guy really worked for free) made me think of how easy would it be in those days to solve the problem of paternity.
With some collaboration between an association of photographers (AF) and a couple big agencies this could be done pretty fast.
I think of a notarized system which can be pretty automatized.
First, let's say a photographer goes to an office of the Association of Photographers on a day when a notary is present (which can be once a month for instance). He shows his documents or otherwise prove his identity to the notary. He generate a cryptographic key pair (of the PGP / GPG kind, or anyway any asymmetric cryptography based on open standards) in presence of a notary.
The public key is given to the notary and to a server kept by the AF with some form of notary force of evidence. The private keys remains in the sole hands of the photographer.
When the photographer executes a job, before sending the pictures (RAW, TIFF, JPEG whatever, provided it is a standard which supports IPTC information) he signs the picture and attaches the signature using a well defined IPTC field (or a field to be created anew in the next version of the standard), when I say "signs" I mean he generates a signature which uses an hash according to a certain asymmetric cryptography algorithm, with a program that every photographer will have on his computer as a basic standard tool (it could well be incorporated into editing programs).
At giving the work to the client, immediately before or maybe immediately after, as soon as possible, the photographer connects to the AF server and registers the picture as his own, by trasmitting the signature with the hash.
So on the notarized server there is a signature which mathematically can only have been produced by that picture and by that photographer and on a date certified by the notarized server (asimmetric cryptographic is used on the server to ascertain that the picture was actually sent by that photographer, the notary certifies that that name/public key actually belong to that actual person).
The AF computer does not receive any image, only hashes. Hashes have a very small size and you can keep any number of them on a cheap system.
The photographer finds later on that the client is not paying and is forced to sue him.
He can show the TIFF picture to the judge, who will agree is the picture used by the client, but would until now be in doubt whether the suing photographer is actually the copyright owner for it.
No doubt any more! The described informative system can prove that on day X was deposited in the AF server a signature, which was inserted in the AF server on date X, and which produces a hash which is actually the hash of that picture with a signature that can only have been produced by that photographer.
Insolvency apart, if the client is in good faith but has lost track of whose the photo is, he can connect to the AF server, give it the signature of the picture, and have the name and contact information of the photographer who registered the picture. The server would always have updated information, so the photographer can be contacted even after having changed address many times (the update would require to see the notary).
If a quite inhonest person (QIP) steals the image, modifies just a bit of it (so that the hash changes) and registers it on the AF server with his own signature, the picture (the hash with signature) will be registered as belonging to QIP, but when you sue him, you can show the judge a picture (just as you would show him a slide) with a notarized answer from the system stating that the picture has been registered by YOU on day X time Y (QIP has registered it later than you, if you have registered it fast).
The costs of the system would be very small: professional photographer in any given nations are not in the zillions, each of them would easily pay some tenth of dollar per year in order to finance the system. Photographic Agencies would certainly adhere and finance the system also.
The entire technology to do so exists since 15 years at least at computer desktop level, and is extremely efficient. I created my first PGP key in 1996 and I could generate a 1024 bit signature with a 30 MHz Amiga in a few seconds, a hash from a 100MB picture might be longer to calculate but who cares, fast computers of our days can crunch an enormous computational work in a minute. Also the hash can be computer by taking only certain pixels of the picture.
IPTC fields already exist and are supported by the great majority of graphic formats. Maybe a new field can be created on purpose, or a generic info field can be used.
The only real costs would be related to the notarization process.
I wander why such a thing has not being realized yet. Just to make an example, Corbis registers a copyright for every picture they represent. So the costs of the actual systems are high (whether you want to register copyright or not). An automated system would generalize copyright registration and keep costs very very low.
What do you think of such a system?