Skip to content

Ninth Circuit Revises UMG Recordings v. Shelter Capital Opinion

Veoh logoI want to briefly interrupt the hubbub over Kirtsaeng (which I fully intend to contribute to at some point) to note that last week the Ninth Circuit released an updated opinion in UMG Recordings v. Shelter Capital Partners, the Section 512 immunity case it decided back in December 2011. The old opinion, 667 F.3d 1022 (9th Cir. 2011), is now vacated. Apparently the panel conducted a rehearing and the new opinion modifies the old one in a few ways, primarily to bring the opinion in line with the Second Circuit’s opinion in Viacom v. YouTube, 676 F.3d 19 (2d Cir. 2012). I’ve created a redlined version showing the changes (aside from minor differences like page numbers).

That’s a pretty interesting development, I think. I can’t think of another copyright case where an appellate court changed its opinion after having been persuaded by the reasoning in another circuit’s opinion. Certainly it would be better if the Ninth and the Second Circuits “talked to each other” more often in this way.

In any event, the outcome is still the same: UMG lost on all of its arguments that Veoh didn’t fall within the safe harbor, and that Veoh’s investors were secondarily liable for infringement; Veoh still loses on its argument that its Rule 68 offer of judgement should get it attorneys fees. But there are several important alterations to the reasoning. First, the amended opinion removes the problematic inclusion of a knowledge element in the Section 512(c)(1)(B) exception for ISPs that “receive a financial benefit directly attributable to the infringing activity, in a case in which the service provider has the right and ability to control such activity.” The prior opinion had held that the ISP must be aware of infringing material in order to control it. The problem, as the Second Circuit pointed out in Viacom, is that that requirement collapses the 512(c)(1)(B) exception into the 512(c)(1)(A) exception for knowledge or awareness. Awareness of infringing material removes the 512 immunity under either the actual knowledge or the red flags provisions; 512(c)(1)(B) would never apply in a case where 512(c)(1)(A) didn’t already remove the immunity.

There is still an issue here, which is that it is not clear how to interpret 512(c)(1)(B). It looks a lot like an exception for vicarious liability; that is, it looks like it excludes vicarious liability from 512(c) immunity. But given the breadth with which courts have applied the vicarious liability test, that would make little sense. As several courts have noted, 512(c) presumes that an ISP can remove material from its system, which has ordinarily been sufficient to give it the “right and ability” to control conduct for purposes of vicarious liability; and direct financial benefit has been interpreted to include any activity that draws visitors to the venue. Almost every ISP would face vicarious liability under such a standard, and would lose its 512 immunity at the same time, making 512(c) pretty much useless.

The Second Circuit in Viacom struggled to find a way around this, departing briefly from its strictly textualist interpretation of the statute. “Something more” must be required to lose immunity under the 512(c)(1)(B), it concluded; but it could not determine exactly what it was, and remanded to the district court to figure it out. (This is the same district court that largely punted on the 512 issues to begin with, so I’m not optimistic for a sophisticated analysis.) The Ninth Circuit in the revised UMG opinion takes the same tack, following the Second Circuit’s suggestions that the ISP could demonstrate the necessary “something more” by exerting substantial control over the content of the website, or by intentionally inducing the infringement. But those suggestions strike me as simply word games; they are both ways of reading in a knowledge-substitute into the vicarious liability concept. Substantial control over the content of a website, including the infringing content at issue, would at least give the ISP reason to know of the infringement; and inducement is a variety of contributory infringement that requires intent. The problem is that 512(c)(1)(B) does essentially mirror the test for vicarious infringement. The problem is that the test for vicarious infringement is far too broad. That can’t be easily fixed through a textualist interpretation of 512.

The second difference between the old and the new UMG opinions concerns the “knowledge” exceptions to 512 immunity, which abrogate ISP immunity when the ISP has “actual knowledge” of infringement or is “aware of facts or circumstances from which infringing activity is apparent,” the so-called “red flags” test. The distinction between these two exceptions has never been entirely clear. The Second Circuit in Viacom concluded that the difference between the two exceptions was not that one of specific versus general knowledge, but rather subjective versus objective knowledge. Actual knowledge is founded on a belief actually held by the defendant ISP. “Red-flags” knowledge, according to Viacom, is knowledge of facts or circumstances from which a reasonable person would conclude that they know of a fact, regardless of whether the defendant in fact drew that conclusion. It is what in torts is called “reason to know” of a fact. In the revised UMG opinion, the Ninth Circuit adopted this distinction. Both courts now hold that mere knowledge that the system is capable of hosting, and likely does host somewhere, infringing material will not give an ISP objectively reasonable “reason to know” of infringement.

Third, with respect to whether Veoh fell within the “knowledge” exceptions to 512 immunity, which abrogate ISP immunity when the ISP has “actual knowledge” of infringement or is “aware of facts or circumstances from which infringing activity is apparent,” the panel dropped a new footnote after a quotation from Corbis Corp. v. Amazon.com, Inc., 351 F. Supp. 2d 1090, 1107 (W.D. Wash. 2004). The quotation noted that in that case, as in UMG, the copyright owner had opted not to send any takedown notices under Section 512, which “stripped it of the most powerful evidence of a service provider’s knowledge — actual notice of infringement from the copyright holder.” The new footnote attempts to backtrack somewhat from the idea that notices provide knowledge:

Notably, the statute specifies that notice of infringement by or on behalf of a copyright holder that does not substantially comply with § 512(c) “shall not be considered . . . in determining whether a service provider has actual knowledge or [has red-flag knowledge].” 17 U.S.C. § 512(c)(3)(B)(i). Proper DMCA notice under 17 U.S.C. § 512(c)(3) provides only a claim of infringement, and is not necessarily sufficient by itself to establish actual or “red flag” knowledge. Instead, proper DMCA notice gives rise independently to an obligation to remove the allegedly infringing material as well as to procedures for ascertaining whether the material is indeed infringing. See § 512(g).

It’s difficult to be certain, but the panel appears to be attempting to head off an argument — not relevant in this case, since there were no notices — that compliant notices of infringement of a particular copyrighted work necessarily gives rise to knowledge or awareness that other copies of that work on the system are infringing. (Non-compliant notices, as the statute states, have no effect whatsoever.) This sort of claim was a feature of the Viacom case, and the language from Corbis that was quoted in the original panel opinion appears to state that such notices constituted “the most powerful evidence” of the requisite knowledge that would lose an ISP its immunity. The new footnote qualifies that by stating that such notices are “not necessarily sufficient” by themselves to constitute knowledge or awareness of infringing activity; the only certain effect is to obligate an ISP to pull down the identified material if it wants to preserve its statutory immunity.

Finally, the modified opinion adds a paragraph agreeing with the suggestion in Viacom that “willful blindness” should count as knowledge or awareness of infringing activity, but finding that it was not present in this case. The Viacom court largely saved that issue for another day, and the revised UMG opinion preserves the issue for later as well.

3 thoughts on “Ninth Circuit Revises UMG Recordings v. Shelter Capital Opinion”

  1. I think you hit the nail on the head with vicarious liability – the statute clearly incorporates the common law standard. The real problem is that standard might kill the safe harbor – one need only reread Fonovisa to relive the heart attack that is vicarious liability. Maybe the “direct” part of financial benefit can be read more narrowly, though. Or maybe the “control” prong requires a practical ability to do so…

  2. Vicarious liability is a rabbit hole. I’ve been digging into this lately, attempting to figure out how contributory and vicarious liability came to be in their current states, and vicarious liability is deeply unclear. It seems to slip away into contributory liability even in the old cases — even the dance hall cases are arguably about contributory liability. (You hire a band and tell them to play some music; is it really going to be beyond your knowledge when they start performing copyrighted tunes?)

    I think one place where things may be going awry is that “right and ability to control” shouldn’t be about the activity; it should be about the *actor*. HL Green had that in the Shapiro Bernstein case, but if control over the actor is lacking, then the “something more” you need is knowledge, but that’s because it’s contributory liability that needs to be shown, not vicarious. That’s my tentative theory for the moment, anyway.

  3. I suspect you may be right, Bruce. I’m teaching contributory vs vicarious liability at the moment, and it seems to get harder and harder to explain the difference because knowledge always seems to come into it, and the results of the digital cases at least seem to be the same whichever theory you apply precisely because you always end up talking about knowledge of some kind.

Comments are closed.