User-Generated Content Mush

Some major copyright holders (though notably not the record companies) and some user-generated content sites (though notably not YouTube) have signed a bunch of “User-Generated Content Principles.” The basic idea seems to be that if the sites first rip out their own intestines and then use “Identification Technology” to filter out and block most infringing stuff from showing up on their sites in the first place, and finally promptly take down any infringing stuff that should happen to slip through upon notice, the content conspiracy won’t show up with lawyers and Lugers. In other words, voluntarily go well beyond what copyright law requires of you and we’ll pretend that we’re letting you off out of the warmth of our stony little hearts, rather than because we’ve got no case.

For me, the real “and a pony” moment comes in principle 3:

UGC Services should use effective content identification technology (“Identification Technology”) with the goal of eliminating from their services all infringing user-uploaded audio and video content for which Copyright Owners have provided Reference Material (as described below). To that end and to the extent they have not already done so, by the end of 2007, UGC Services should fully implement commercially reasonable Identification Technology that is highly effective, in relation to other technologies commercially available at the time of implementation, in achieving the goal of eliminating infringing content.

(emphasis added). That’s not a rule. It’s not a standard. It’s not even a principle. It’s mush. The Identification technology, that wonderful magical animal, should be “commercially reasonable” but also have a goal of “eliminating…all infringing…content.” How well does it need to work? It should be “highly effective,” but only “in relation to other technologies.” Do you know what that means? I don’t.

This is where the bodies are buried, my friends. Sainted Congress itself couldn’t have written a more mealy-mouthed compromise. Copyright holders will claim that whatever technology the UGC sites are using isn’t even “effective,” let alone “highly effective.” The sites will reply that in comparison with the other crud out there, their systems are ust about as effective as you can get. They’re both right. If software that could distinguish infringement from noninfringement automatically and reliably were something a UGC site could just order up, we wouldn’t be in this mess in the first place. Even identification technology that’s extraordinarily effective in relative terms is still ineffective in absolute terms, and I full expect everyone involved to argue whether “highly effective” is an absolute or relative term until the cows come home.

The scary thing is that the RIAA may well have been holding out for more. What’s their endgame—“we get to control everyone distributing, quoting, sampling, or referencing our music—now and forever amen.”