In the abstract, the term could mean any services offered online, including the service of making copyrighted works available to the public. Or the term could mean services speciﬁc to being online (other than network access, for which the deﬁnition speciﬁcally provides). Under the ﬁrst interpretation, anyone who operates a website is a ‘service provider’. Under the second, an entrepreneur who hosts a website is a ‘service provider’, as is one who provides online search services; the entrepreneur who makes content available [here she means “allows users to post content”], however, would not be a ‘service provider’ because the services provided are not Internet-speciﬁc. One can provide content from a variety of platforms (e.g., print, broadcast), but one can host or link to a website only via the Internet.
Jane C. Ginsburg, User-Generated Content Sites and Section 512 of the US Copyright Act, in Irini A. Stamtoudi, ed. Copyright Enforcement and the Internet (Kluwer Law International 2010).
What is the difference between “hosting a website” and “providing content”? What does she think a website is? One can “host” a discussion group in the letters column. Or, run it the other way, it’s only possible to stream video online. Statutory interpretation solely for the purpose of suggesting that YouTube was only accidentally covered by the DMCA, and only reluctantly conceding coverage, is not confidence-inducing. See also: “The examples of service providers given in the House Report consist entirely of enterprises who provide ‘space’ for third-party websites and fora, not the operators of the websites themselves.” Id. True, yet puzzling why this is not a description of YouTube, except for some oddly formalistic definition of “website” where a separate domain name or subdomain is required, which would be inconsistent with the history of concern for protecting the host of discussion groups (cf. Prodigy) and would require us to impute the term “website” into the statute for no articulated reason. Ginsburg basically thinks that service providers like YouTube could make money from infringing content, and so should be outside the safe harbor, but for-profit status doesn’t help distinguish YouTube from Netcom.
She makes an interesting point about the “right and ability to control” portion of the safe harbor. I’m completely unconvinced by her claim that it should be ok to equate “right and ability to control” with “ability to take down after notice,” such that service providers using the user-directed storage safe harbor will always have the right and ability to control the infringing conduct by virtue of the notice-and-takedown system they use to qualify for the safe harbor. She suggests that the financial benefit requirement will mean that some ISPs might still theoretically qualify for the safe harbor (though the trend is to read “direct” financial benefit incredibly broadly). But she points out that, to the extent that her reading makes “right and ability to control” incoherent/redundant, it’s also not clear what the alternative is: given the structure of UGC websites, if “ability to take down” isn’t “right and ability to control,” then what would be? (My answer, for what it’s worth, is that we should treat the concept as it would be treated in ordinary, non-copyright vicarious liability: some sort of agency relationship, not the extremes to which copyright infringement has gone.)