UMG v. Shelter Capital: A Cautionary Tale of Rightsowner Overzealousness

By Eric Goldman

UMG Recordings, Inc. v. Shelter Capital Partners LLC, 2011 WL 6357788 (9th Cir. Dec. 20, 2011). My prior blog posts on district court rulings on Veoh’s 512(c) safe harbor and attorneys’ fees/Rule 68.

Make no mistake, web hosts and their investors got a major 512(c) victory in this ruling. The Ninth Circuit, building on its favorable but convoluted ruling in Perfect 10 v. ccBill, wrote a decisive and clear (well, as clear as the 9th Circuit gets…) opinion interpreting the crucial 512(c) safe harbor. This opinion is so comparatively lucid that I plan to substitute it into my Internet Law reader next Fall as a replacement for the Io v. Veoh and Viacom v. YouTube district court rulings.

But also make no mistake: this case reminds us why we need to strike a fair balance between rightsowners and technology providers, or else our system will break down. This case’s real result is that Veoh is legal, but Veoh is dead—killed by rightsowner lawfare that bled it dry. Meanwhile, rightsowners wrongly assessed the legality of Veoh, but the worst consequence they suffered was overpaying their lawyers. Indeed, UMG isn’t liable under 17 USC 512(f) for sending bogus takedown notices because they never sent any notices at all., nor is UMG liable for Veoh’s attorneys’ fees. UMG’s decision-makers walk away from this car crash, muttering under their breath that the Ninth Circuit misunderstood their brilliant legal arguments, but they still get to go to their cushy jobs tomorrow. The same can’t be said for Veoh, even though it “won.” Veoh’s employees? On the street. Veoh’s investors? SOL. Veoh’s community? Kicked to the curb.

This case outcome—Veoh is legal, but Veoh is dead—highlights one of the many reasons why so many people are so opposed to SOPA/PROTECT-IP. Those proposals don’t make rightsowners fully internalize the cost of their actions, such as the economic losses suffered by erroneously accused targets. Of course rightsowners will overclaim when there’s no real downside to doing so; that’s just human nature. (And please, I don’t want to hear any BS that rightsowners will never get it wrong. See, e.g., Viacom v. YouTube). Without proper calibration, rightsowner overclaiming threatens to wreck the entire Internet ecosystem.

A partial fix to SOPA/PROTECT-IP would make rightsowners bear the cost of their overclaiming. Make them put up a $1 billion bond for the privilege of sending cutoff notices; and pay liberally out of that bond if the rightsowners get the law or facts wrong. Write checks to the investors and employees whose economic expectations are disrupted when rightsowners get it wrong. Write checks to the payment service providers and ad networks who turn down money from legally legit businesses based solely on rightsowner accusations. Heck, write checks to the users of those legit services who are treated as inconsequential pawns in this chess match. Sure, a $1B bond obligation with liberal payouts would turn cutoff notices into a sport of kings that only the richest rightsowners could afford, but perhaps that’s the way it should be. A rightsowner’s decision to send a cutoff notice should be a Big Deal, the equivalent of going to Defcon 5, and not like sending holiday cards to distant relatives you last saw at Ethan’s bar mitzvah.

Unless (until?) Congress wrecks the Internet with SOPA/PROTECT-IP, 17 USC 512(c) still matters a lot to the Internet ecosystem, and this ruling has a lot of good news for web hosts. It’s a long opinion, as 512 opinions usually are. Some highlights:

* Since its passage, 512(c) has had a crucial ambiguity: did it provide a safe harbor for all three flavors of infringement (direct, contributory, vicarious) or just direct infringement? The legislative history was clear that the safe harbor applied to all three flavors, but the literal text of the statute seemed, by its very terms, to exclude contributory and vicarious infringement. Remarkably, most 512(c) cases sidestepped this fundamental interpretive issue. In contrast, this case confronts it head-on and, perhaps for the first time in an appellate ruling, indicates that 512(c) applies to all three flavors of infringement. For example, the court says:

Given Congress’ explicit intention to protect qualifying service providers who would otherwise be subject to vicarious liability, it would be puzzling for Congress to make § 512(c) entirely coextensive with the vicarious liability requirements, which would effectively exclude all vicarious liability claims from the § 512(c) safe harbor….Although in some cases service providers subject to vicarious liability will be excluded from the § 512(c) safe harbor, in others they will not.

To achieve this outcome, the Ninth Circuit says that “right and ability to control”—language that appears in both the standard test for vicarious copyright infringement and as an exclusion to 512(c)’s availability—means different things in those contexts. I presume the same applies to “direct financial interest,” though that didn’t come up here. Notice this is a result only lawyers could love: the exact same words have different meanings depending on whether they are in the plaintiff’s prima facie case or in the defendant’s affirmative defense.

The good news is that the opinion reads “right and ability to control” as dependent on knowing of specific problems that need to be remediated, like the Viacom v. YouTube opinion did:

a service provider must be aware of specific infringing material to have the ability to control that infringing activity within the meaning of § 512(c)(1)(B). Only then would its failure to exercise its ability to control deny it a safe harbor….A service provider’s general right and ability to remove materials from its services is, alone, insufficient. Of course, a service provider cannot willfully bury its head in the sand to avoid obtaining such specific knowledge.

Unfortunately, the last sentence reminds me a little of the gratuitous dicta in Tiffany v. eBay about “willful blindness,” which trademark plaintiffs have already started to mine. Trying to scrounge for any angle, I’m sure copyright plaintiffs will start digging around for evidence that service providers buried their head in the sand, making the most tendentious interpretations of fact that aren’t damning in the least.

* UMG argued that 512(c) only provided a safe harbor for personal cloud storage. Under this reading, the moment a file stored in the cloud was available to third parties, 512(c) dropped away. The court brushed away UMG’s argument, and I can’t believe UMG thought it was worth pressing. The argument made no sense historically (no one was offering personal cloud storage lockers when the DMCA was passed), it contravened numerous provisions in the statute that clearly suggested otherwise, it would have conflicted with a long list of precedent cases that applied 512(c) to public hosting, and it’s almost impossible to construct a fact pattern where a user uploading to a personal cloud locker commits copyright infringement (or that a copyright owner would learn about this storage sufficient to send a 512(c)(3) notice specifying the file’s location).

As a corollary, 512(c) applies even if the service provider slices-and-dices the user-submitted file, such as transcoding the file and extracting metadata. This is consistent with earlier 512(c) cases, but now it’s Ninth Circuit law.

* The court rejects UMG’s many arguments that Veoh had impermissible scienter. UMG’s hubris was insane. For example, UMG argued that Veoh hosted music videos, and that because it didn’t have licenses to the music, its general knowledge categorically disqualifies Veoh from 512(c). The court shreds UMG’s argument, noting that Veoh did have direct arrangements with music video producers, and UMG’s argument (if you host music, you must know you’re infringing) would negate the entire 512(c) safe harbor. Instead, the court emphasizes the importance of the notice-and-takedown scheme in 512(c) because “Copyright holders know precisely what materials they own, and are thus better able to efficiently identify infringing copies than service providers like Veoh, who cannot readily ascertain what material is copyrighted and what is not.” Thus, the court “hold[s] that merely hosting a category of copyrightable content, such as music videos, with the general knowledge that one’s services could be used to share infringing material, is insufficient to meet the actual knowledge requirement.” Nor can such knowledge count as a “red flag.”

The court isn’t any more impressed with UMG’s argument that Veoh bought Google Adwords keyed to the names of UMG artists like 50 Cent, Avril Lavigne and Britney Spears. The court responds:

50 Cent, Avril Lavigne and Britney Spears are also affiliated with Sony-BMG, which gave Veoh permission to stream its videos by these artists. Furthermore, even if Veoh had not had such permission, we recognize that companies sometimes purchase search terms they believe will lead potential customers to their websites even if the terms do not describe goods or services the company actually provides. For example, a sunglass company might buy the search terms “sunscreen” or “vacation” because it believed that people interested in such searches would often also be interested in sunglasses. Accordingly, Veoh’s search term purchases do little to demonstrate that it knew it hosted infringing material.

I wouldn’t rely on the Ninth Circuit for SEM advice, but did the court really say that a site can buy the keyword “Britney Spears” to reach teeny-boppers, even if the site doesn’t offer any Britney Spears music? Hmm…

The court also rejects UMG’s argument that takedown notices from the RIAA should have prompted Veoh to go find additional videos from the same artists mentioned in the takedown notices. As ccBill said, service providers don’t have that affirmative investigatory duty; it remains solely on copyright owners’ shoulders.

Finally, the court rejects UMG’s efforts to dig up old newspaper quotes from Veoh executives acknowledging that their site contained infringing material. The court appropriately notes that 512(c) assumes UGC sites will contain some infringing items. That’s why copyright owners should send takedown notices, and why they shouldn’t bitch if they don’t.

One sour note: The court says that a user-submitted complaint that he’d seen infringing content on the site, and fingering a specific other user, could constitute a “red flag of infringement” even if the user complaint didn’t constitute a 512(c)(3) notice. UMG didn’t make any progress here based on the specific facts, but unfortunately the Ninth Circuit opened up the door for copyright owners looking for notifications from non-copyright owners that the copyright owners can turn into red flags. Of course they are going to find such notices; what UGC site hasn’t gotten dragged into intra-user disputes? Unfortunately, 512(c) discovery is already ridiculously expensive, and hunts for needles in the haystack like this only add to everyone’s costs—with almost no payoff because sites should be free to ignore user gripes (non-512(c)(3) notices) in their considered judgments. Until the Ninth Circuit fixes this in a future opinion, this sloppy discussion means UGC sites must address non-512(c)(3) gripes about potential copyright infringement at peril of being accused of having red flags of infringement. This isn’t what Congress intended, so it’s a bummer the otherwise-solid opinion went off the rails here.

* The ruling absolved Veoh’s investors of liability. In a footnote, the court recognizes the importance of keeping investors free of liability, especially when the site actually qualifies for the 512(c) safe harbor:

Congress was no doubt well aware that service providers can make the desired investment only if they receive funding from investors like the Investor Defendants. Although we do not decide the matter today, were we to hold that Veoh was protected, but its investors were not, investors might hesitate to provide the necessary funding to companies like Veoh, and Congress’ purpose in passing the DMCA would be undermined.

The court says that UMG isn’t arguing that funding a venture is enough to create liability; the investors must be involved with the business operations. The court tries not to overrule one of the Napster district court rulings regarding investor liability by saying that, in the Napster case, only one investor (Bertlesmann) was involved with Napster. So long as there are more than 1 investor—and frankly, when won’t that be the case?—each individual investor can’t have sufficient control to trigger liability. In response, UMG tried to argue that Veoh had three outside board members from investors and they collectively controlled the board, but the court said UMG didn’t adequately allege that they were in cahoots with each other.

It would have been so much better if the court had just rejected investor liability outright rather than nose-counting board seats and agreements among board members. Expect copyright owners to impose discovery heck on investors, looking for any evidence that smacks of coordinated efforts among them, and expect rightsowners to make mountains out of molehills like stockholders’ agreements. Nevertheless, the Ninth Circuit opinion has enough language to raise the bar on investor lawsuits that district courts should toss these efforts on summary judgment (the needle-in-haystack discovery hunts are going to make motions to dismiss hard). Let’s hope the district courts set the bar high enough that copyright owners eventually get discouraged in pursuing investors.

* The ruling basically eliminates FRCP 68 for copyright cases (and presumably any other statutes that have express fee-shifting provisions). Rule 68 says that if a defendants offers a jdugment, is refused, but then achieves final results better than the settlement offer, the plaintiff must pay all attorneys’ fees after the offer. The idea is to motivate plaintiffs to accept fair settlement proposals—they have to be so confident they’ll do better the settlement offer because they have to pay off the defendant if they are wrong. Rule 68 provides a useful tool from a game theory standpoint, but the court eviscerates it for copyright cases. The court says that Rule 68 offers can pay off only if the judge chooses to award attorneys’ fees under 17 USC 505. No 505 fee shift, no Rule 68 fee shift. But, why is Rule 68 needed if the judge has to make a 505 fee award anyway? This makes no sense. Rebecca discusses this issue in more detail.

I know a lot of folks are interested in how this case affects Viacom v. YouTube. The news is all good for Google. First, until the Second Circuit issues its opinion, this opinion is the new high water mark for 512(c) cases; and it is the governing law of the Ninth Circuit. Second, I imagine the Second Circuit panel will take a look at this opinion, and they may choose to defer to it. The opinion isn’t airtight analytically, but it’s persuasive enough. Third, in the unlikely situation that Google loses in the Second Circuit, this opinion sets up the possibility of a bona fide circuit split that might open the door for a Supreme Court appeal.

Other coverage of the case:

* Techdirt

* Rebecca (focusing on Rule 68)


* Michael Barclay