You've oversimplified the DMCA. The "safe harbor" you're referring to includes the requirement that the provider "does not have actual knowledge that the material or an activity using the material on the system or network is infringing; [or] in the absence of such actual knowledge, is not aware of facts or circumstances from which infringing activity is apparent".
You have to be able to say with a straight face that you couldn't reasonably have known the material was infringing a priori to be eligible.
> You have to be able to say with a straight face that you couldn't reasonably have known the material was infringing a priori to be eligible.
Copyright infringement is a matter of permission. I'm not sure that it's even possible, a priori, to say that certain material is infringing, unless you have personal knowledge of who uploaded it and who is authorized to. You can usually make a good guess; but it's just a guess.
This even more complicated thanks to the fact that the record labels have authorized marketing agents to release "unauthorized" releases, even ones marked as leaks, via independent channels. They've even have had to un-DMCA them when their own legal departments took them down.
Proof of that statement can be found in the court filings for the Viacom v. YouTube case, incidentally. Lest you think that this was just the assertion of YouTube's lawyers, I will point you further to the fact that Viacom's own lawyers were forced to agree to remove their claims of infringement after it was proven that Viacom itself had uploaded those particular files.
And that? That happened twice. That's after Viacom's own expensive lawyers (allegedly) performed due diligence.
In short, I would like to point out that while you can have a pretty good idea what content is and is not authorized, there are many good reasons, based on actual events, that could give a person a reasonable doubt about whether something might be authorized, especially if there's no DMCA complaint. I can give you a long list incidents if you like, just from memory.
My point is that even the rights holder might not know if they've authorized it, and they're the only people who can be expected to know who all they've authorized.
It's an interesting point because knowing/not knowing is not always black and white. While it's almost impossible to determine 100% whether or not something is infringing, not collecting enough data to tell does NOT work. The courts aren't that interested in the specifics, it's "are you the Pirate Bay or are you Facebook?"
You have to be able to say with a straight face that you couldn't reasonably have known the material was infringing a priori to be eligible.