Veoh Gets Yet Another Terrific 512 Defense Win–UMG v. Veoh
By Eric Goldman
UMG Recordings, Inc. v. Veoh Networks, Inc., 2:07-cv-05744-AHM-AJW (C.D. Cal. Sept. 11, 2009)
What’s the difference between the market leader and an also-ran? When the also-ran wins its third big legal victory in a row, the first question everyone asks is “what does this mean for the market leader?”
Specifically, when the copyright infringement lawsuits were filed against YouTube in 2007, everyone got really excited about a looming legal determination of the legitimacy of YouTube’s practices. (See, e.g., 1, 2). But, we haven’t learned very much about substantive law from YouTube’s litigation in the 2+ years since. Instead, while the YouTube cases are mired in procedural heck, Veoh has notched three powerful defense-side wins that surely help YouTube (see my previous posts on the Io case and the earlier 512 ruling in this UMG case). These cases have established Veoh as the litigation leader of the user-generated online video business, even if it’s not the market leader.
In the December ruling in this case, Judge Matz concluded that the videos in Veoh’s network were “stored at the direction of users” even though Veoh’s system technologically manipulated the user-supplied video. Having established that, this ruling addresses whether Veoh met the other requirements of 17 USC 512(c).
Actual/Red Flags Knowledge
A service provider loses 512(c) eligibility if it does not expeditiously remove infringing items after getting actual knowledge or has sufficient awareness of infringement. This standard has proven tricky because plaintiffs keep trying to establish knowledge and awareness based on something other than the 512(c)(3) notices contemplated by the statute. In this case, everyone agreed that Veoh responded expeditiously to 512(c)(3) notices, so UMG argued—and the judge rejected—that Veoh got enough culpable scienter from other things, such as
* Veoh was in the music business, and music is protected by copyright.
* Veoh had tagged videos as “music videos”
* Veoh bought search ads keyed to the titles of UMG copyrighted works (but all identified instances were for artists where Veoh had a license covering some of the artists’ works)
* a non-512(c)(3) compliant notice categorically opting out for specified artists. The court rightly rejects such notices as construing actual knowledge because artist names are too imprecise (they do not identify either the infringed works or the infringing works.
The latter point is a contentious one. Compare the wrongly decided 2001 ALS Scan v. RemarQ case, which said that the service provider had to assume that everything published in USENET newsgroups containing the plaintiff’s trademarks was categorically verboten. Also compare the Hendrickson case (distinguished in a footnote). However, the judge leaves open the possibility that a non-512(c)(3) compliant notice could still confer actual knowledge if the copyright owner categorically identifies the title of the infringed work or made the representation that all works of a particular artist were unconsented.
The judge also rejects the argument that Veoh had generalized knowledge that widespread infringing activity was occurring on its site.
Finally, UMG argued that Veoh got the requisite scienter from its failure to use filtering technology properly. Veoh went through three iterations of filtering: first, none, then in 2006, a hashing technology, and finally in 2007, Audible Magic. UMG argues that Veoh should have adopted Audible Magic earlier and should have filtered its back catalog quicker once it adopted the technology. The court rejects this argument because (1) Veoh wasn’t obligated to adopt filtering technology at all, and (2) it showed good faith by migrating from a poor solution (the hashing technology) to a more robust system.
To be eligible for 512(c), a service provider also must show that it lacked the “right and ability to control” the infringements. The court clarifies that this must mean something other than the operation of the system and the ability to take down content; otherwise, every service provider would be disqualified. Veoh’s ability to deploy filtering software is also irrelevant per 512(m) (a rarely discussed corner of the statute).
However, this court is bound by the awful and antiquated Napster precedent, which had a very expansive view of “right and ability to control.” The court sidesteps Napster by saying the Ninth Circuit’s Napster interpretations only apply to the prima facie case of vicarious copyright infringement, not the identical words contained in 512(c). This is an awkward conclusion because (1) the words are identical, and (2) in ccBill, the Ninth Circuit held that the 512(c) words “direct financial benefit”–immediately adjacent to “right and ability to control” in 512(c)–should be interpreted the same as the usage in the prima facie case. So this court reaches the tough conclusion that “direct financial benefit” means the same thing in 512(c) and the prima facie case while the words “right and ability to control” do not.
This dichotomous language parsing seems particularly unstable. However, it is a logical consequence of the Ninth Circuit’s steadfast but ill-fated refusal to overturn the analytically corrupt Napster ruling, which has been horking Ninth Circuit Internet copyright jurisprudence since 2001.
Termination of Repeat Infringers
Finally, the service provider is obligated to terminate repeat infringers. UMG argued that Veoh is disqualified because it doesn’t automatically terminate users whose videos are filtered out by the Audible Magic technology. The judge disagrees because Audible Magic’s system is a black box, so no one knows how many false positives it produces. Further, other courts have said that a compliant 512(c)(3) notice isn’t accurate enough to automatically trigger an obligation to kick the target user off, so Audible Magic’s fallibility makes its assessments even less compelling.
UMG also argued that it was unreasonable that Veoh did not kick off users who had uploaded multiple files that were covered under the same 512(c)(3) notice. (Veoh had a two-strikes rule for those users). The judge rejects that argument too, saying that Veoh’s policy was appropriate.
Overall, this case is another terrific defense win. In every instance of ambiguity, the judge sided with the defense, repeatedly noting that copyright owners have to shoulder more burden to prepare compliant 512(c)(3) notices and service providers have no burden to filter or take other steps that copyright owners demand but aren’t in the statute. Some of this categorical siding with the defense might also reflect that UMG appears to have annoyed this judge. I laughed out loud when I saw FN 10, which says “Veoh cited Io Group extensively in its opening brief. UMG did not mention it at all in its Opposition.” You can run from precedent, UMG, but you can’t hide.
While there is a lot of good news in this opinion, it’s also noteworthy for what it didn’t say. As I’ve mentioned repeatedly in past blog posts, 512(c) has a critical embedded ambiguity–does it insulate service providers from all three flavors of civil copyright infringement (direct, contributory and vicarious) or only direct infringement? Amazingly, courts generally have avoided discussing this crucial threshold consideration, instead proceeding ahead with an assumption (not always the same assumption from case to case!) on this crucial question but not resolving it. This case is no different; the court apparently makes an unexamined assumption that 512(c) is a defense to all three flavors of infringement. This is a terrific assumption for Veoh, YouTube and all 512 defendants, but it would be even better if the court actually said that.
The court does come closer than many other courts to addressing this ambiguity when it said there was a difference between the prima facie vicarious copyright infringement and 512(c) uses of the phrase “right and ability to control.” However, I will be very interested to see if the Ninth Circuit agrees with that dichotomy on appeal. Even if they affirm the judgment, I suspect this dichotomy won’t survive intact.