Part 2 of 2: The Second Circuit Punts on Third Down

Last time, I laid out the context for the Second Circuit’s decision in the Viacom v. YouTube case–i.e., the state and open issues of the law of the DMCA safe harbors. This time, I’ll get into what the Second Circuit actually said, pointing out where the Second Circuit agrees with, and diverges from, the Ninth Circuit’s reasoning in Shelter Capital v. UMG (the “Veoh case”).

Knowledge Requirement Is Limited to Specific Knowledge

The Second Circuit agrees with the Ninth Circuit that only knowledge of specific acts of infringement may defeat the Knowledge Requirement, whether under the actual knowledge or “red-flag” prongs.

The Second Circuit goes a bit further and describes “red-flag” knowledge as “objective” knowledge, as opposed to subjective knowledge. I.e., a reasonable person would have had knowledge (without conducting an investigation!) regardless of his or her actual knowledge. If that’s the case, the scope for red-flag knowledge must be very narrow indeed, since a person with so much awareness would almost always have enough facts to constitute actual (subjective) knowledge.

As it happens, the Second Circuit found instances that arguably show actual knowledge, in the form of emails from YouTube executives that at least suggested they knew that specific files were likely infringing. E.g., “[T]he CNN space shuttle clip, I like. [W]e can remove it once we’re bigger and better known, but for now that clip is fine.” The court did not say that the emails proved that the executives had actual knowledge, only that there was enough doubt that a jury should be the one to figure it out. This accounts for some of the remand. But the scope of the remand is very limited: only for the “handful” of files actually mentioned in those emails.

Every Time I Think of Infringement, I Go Blind

Then the Second Circuit throws a curve ball, asking the trial court to examine the issue of “willful blindness.” The court acknowledges that the DMCA relieves operators of any obligation to investigate but concludes that willful blindness isn’t the same thing as monitoring or investigating. It’s more like “looking the other way” when you know or strongly suspect that if you looked, you’d see something you wish you hadn’t seen.

How is this different from actual knowledge or red-flag knowledge? Wouldn’t you need to have those types of knowledge before you even triggered willful blindness? The knowledge threshold must be somehow lower for willful blindness, so long as it is combined with some sort of “conscious avoidance.”

A better formulation might be the Seventh Circuit’s in In re Aimster: “knowing or strongly suspecting that [one] is involved in shady dealings [and] tak[ing] steps to make sure he does not acquire full or exact knowledge of the nature and extent of those dealings.” The Seventh Circuit cites as examples criminal enterprises which were designed by the defendant to shield the defendant from actually know what’s going on, i.e., implementing a system that the designer knows is going to result in illegal activity but which is specifically designed to insulate the designer from any actual knowledge of the illegal activity.

In any case, the application must still be quite limited in light of (1) the anti-investigation provision of the DMCA and (2) the prohibition against end-runs around the notice-and-takedown procedures. Willful blindness shouldn’t apply simply because there are so many “red flags” that the operator should have investigated because that would implicate the anti-investigation provision. In addition, the “strong suspicion” must be derived from somewhere other than the rights holder themselves, since they are stuck with the notice-and-takedown regime.

The only example I can think of would be someone like Fung, from Columbia Pictures v. Fung. In that case, the court found that Fung had “red-flag” knowledge because there was so much obviously copyrighted material on his sites that, to avoid knowledge of infringing activity, he would have had to “engage[] in an ostrich-like refusal to discover the extent to which their systems were being used to infringe copyright.” Perhaps Fung would have been better analyzed under the willful blindness standard? Fung was all but asking his users to upload infringing content, but he made sure there no way to prove that he had direct knowledge of the infringement. It’s a bit like musing out loud in a room full of hit men that you so wished someone dead that you’d gladly reward the someone’s killer with $1 million, then pretending surprise when the someone winds up murdered shortly thereafter.

Anyway, we’ll see what the trial court does with this legal football, and then what the Second Circuit does with it when it gets punted back up to it.

I Know Control When I See It, and it’s Not Vicarious

The Second Circuit agrees with the Ninth Circuit that the Financial-Benefit Requirement is not coextensive with the common-law doctrine of vicarious liability.

However, the Second Circuit holds that “right and ability to control” means neither the ability to stop uploads nor the ability to stop specific instances of infringement. In rejecting the latter interpretation, the Second Circuit directly disagrees with the Ninth Circuit.

At a very abstract level, the Second Circuit’s reasoning seems sounder than the Ninth Circuit’s. Under the Ninth Circuit’s reasoning, you need to import the Knowledge Requirement into the Financial-Benefit Requirement, but if you do that, there’s no point in having a Financial-Benefit Requirement. At the same time, both courts agree that the mere ability to prevent uploads isn’t sufficient to defeat the Financial-Benefit Requirement, since that would take the very types of websites Congress sought to protect right out of DMCA safe harbor.

But the Second Circuit isn’t itself very certain about “right and ability to control” means. The only hint it provides is that, in cases where “right and ability to control” have have been found, the providers exercised “substantial influence on the activities of its users.” But it doesn’t provide much else for the district court to go on during remand.

It does cite a (single) district court decision–out of the Ninth Circuit, no less–as the only example of a defendant found to have exercised the requisite “right and ability to control.”* In that case, Perfect 10 v. Cybernet Ventures, a purveyor of nudie pictures sued an “age verification service” (i.e., it tells porn sites whether a given user is old enough to use the site).** Cybernet was not, however, paid for that service by its porn affiliates. Instead, it made money by collecting subscription money from users, who were then given the right to peruse certain sites in Cybernet’s network of porn sites (presumably giving the affiliates a cut). Cybernet was thus financially dependent on the quality of the porn sites–the better the porn site, the more users will subscribe. Cybernet made sure that the porn sites met certain requirements relating to formatting and content, and told its affiliates that it would monitor compliance with these requirements. The district court, in granting a preliminary injunction, found that Cybernet exercised enough control over the affiliate sites to fail the Financial-Benefit Requirement, when coupled with the obvious direct financial benefit.

* It also suggested that the Supreme Court’s Grokster decision might also present a situation where the operator exercised sufficient “right and ability to control.” But I’m not really clear why. I’ve heard it suggested that Grokster-level inducement (which, remember, was based on a lot of bad facts) is so inconsistent with the DMCA safe harbor that the safe-harbor shouldn’t be available to such providers. But Grokster itself sound more like someone who was willfully blind than someone with “substantial influence” on its users.

** The Second Circuit describes Cybernet’s control as having “instituted a monitoring program by which user website received ‘detailed instructions regard[ing] issues of layout, appearance, and content’.” Obviously, there was a lot more to it than that. Worse, the Second Circuit quoted from the wrong part of the opinion! It quoted from the discussion of Cybernet’s liability for vicarious liability, rather than from the discussion of “right and ability to control.” That’s the jurisprudential equivalent of a Freudian slip!

It’s a little bit difficult to import Cybernet Ventures into the context of YouTube and other providers of user-generated content. Reading Cybernet Ventures, one gets the sense that it’s pretty close to the floor of what would constitute “right and ability to control.” If so, the “right and ability to control” will be construed fairly narrowly. It appears to require that the operator substantially participate in the shaping and selecting of content, not merely influencing the content. For the sake of consumers of online content, it is hoped that website operators will still have the right and ability to exclude content based on subject-matter or quality. I.e., operators will be able to locate and delete content that operator finds disgusting, sacrilegious or boring, or content that is too grainy or otherwise technically deficient. Otherwise, YouTube and similar sites would become pretty horrible to use!

In the end, the Second Circuit punts the whole issue of “right and ability to control” back to the trial court for further development of the record, but without giving the poor trial court* much guidance about how to apply those facts (or even what facts should be developed). Again, we’ll see what the trial court does with it, and then what the Second Circuit does when the issue is punted back.

* I’d feel worse, but honestly, the trial court’s summary-judgment order and opinion wasn’t all that great. 

Miscellaneous Issues

Finally, the Second Circuit agrees with the Ninth Circuit’s broad reading of the threshold requirement for § 512(c) safe-harbor protection: “by reason of storage at the direction of a user.” Nothing exciting there.* However, there is one activity that the Second Circuit decided probably didn’t come in under the § 512(c) safe harbor: the third-party syndication of content selected by YouTube. However, it appears that none of the content of which the rights holders complained was part of the syndication scheme, so no harm, no foul.

* I criticized UMG for spending so much time on this issue in the Veoh case, but I’m not as critical here. This is because, in Veoh, there was already Ninth Circuit authority directly on point. The chances of overturning that authority was remote. By contrast, there was no authority on the issue in the Second Circuit, so the arguments had a better shot. I still think they were complete long-shots, though, but Viacom must have thought the reward was worth the risk.

Like the rights holders in the Veoh case, the rights holders in YouTube didn’t press the Repeat-Infringer Requirement. However, some (but not all) of the plaintiffs made what seems to me a pretty silly argument.* You see, YouTube lets certain rights holders have special access to internal tools that would allow rights holders to identify infringing content (and remove it, too). Everyone else has to use YouTube’s publicly-available search feature. This is actually a pretty common arrangement. It is basically argued this arrangement was discriminatory, and that unless everyone has access to those internal tools, YouTube failed to implement a repeat-infringer policy.

* I note that Viacom, the most sophisticated of the plaintiffs, didn’t sign on to this argument.

How in the world would this implement the Repeat-Infringer Requirement? Well, there are actually two prongs to the Repeat-Infringer Requirement. The first, of course, is to implement a reasonable repeat-infringer policy. The second is to “accommodate[] and … not interfere with standard technical measures.” This second prong has been interpreted to mean that an operator can’t prevent you from searching the site for acts of infringement.

In any event, the rights holders apparently didn’t bother to explain how the internal tools are “standard technical measures” and rejects the argument. I’ll add that over in the Ninth Circuit, this provision is interpreted to require only that the rights holder may access and search the operator’s system for infringing content.


As far as I know, there aren’t any further DMCA safe-harbor issues pending before any appellate courts, so we might need to wait for the Second Circuit to re-visit YouTube. If YouTube settles before then, we might be left with something of a circuit split, with rights holders flocking to New York (and away from California) to bring their copyright actions. There is a remote, but not negligible, possibility that YouTube will appeal the decision to the Supreme Court based on the circuit split.

Thanks for reading!