Part 1 of 2: Second Circuit on DMCA Safe Harbor: It’s Complicated
Well, if you were hoping that the DMCA safe-harbor law would clear up with the Second Circuit’s long-awaited opinion in Viacom v. YouTube–that we’d get to the point where folks would know the contours of the safe harbor without having to consult with a lawyer–then last Thursday was, indeed, as Prof. Goldman put it, a “bummer.” There were things to criticize in last December’s Ninth-Circuit decision in Shelter Capital v. UMG (better known as the “Veoh case”), and certainly rights-holders were unhappy with it, but at least you knew where things stood. The basic lesson from Veoh was: comply with the DMCA notice-and-takedown regime, watch out for notices of infringement by non-rights holders, and things’ll probably be OK.
But YouTube muddies the waters–at least, for a while. That’s not really meant as a criticism. Simplicity and “bright line rules” are nice because it saves business folks and consumers money (fewer legal fees) and worry. But an appellate court’s job is, in this case, to interpret a statute. Sometimes the best interpretation is also a complex or “fuzzy” one. True, lawyers are the main beneficiaries, but appellate courts don’t care about that. They’re just trying to get it right.
The Basics of Safe Harbors Under the DMCA
I picked a pretty good time to start blogging about the DMCA safe harbors back in August.* Back then, I was just trying to introduce the concept, through the case against Grooveshark that was brought here in Nashville, which I did in two posts. I explained that there were four major types of safe harbors under the DMCA (five if you count a minor one), and the most important one is found at Section 512(c) of the Copyright Act. Under that provision, an operator of a website (or other computer system) that permits (even encourages) users to upload content is not liable for claims of copyright infringement relating to that content (the actual language is “by reason of the storage at the direction of a user”), if four things are true:
- The operator has an implements a policy repeat infringers (the “Repeat-Infringer Requirement”);
- The operator is genuinely unaware that the content is infringing, either through actual knowledge or “red-flag” knowledge (the “Knowledge Requirement”);
- The operator does not benefit financially and directly from the infringement, to the extent the operator can control the activity leading to the infringement (the “Financial-Benefit Requirement”); and
- The operator expeditiously removes content subject to a proper “DMCA take-down notice” (the “Takedown Requirement”).
* Only, this blog isn’t really about the DMCA, at least, not exclusively. But the DMCA safe-harbor provisions have come to dominate the blog a bit because (1) they’re complex, important and frequently misunderstood (so I wanted to make sure everyone understood them), and (2) there have been now two major appellate decisions on the contours of the most important DMCA safe harbor.
Unfortunately, even though the DMCA is now about 12 years old, there’s lots we don’t know about these four requirements. For example:
- Does the one providing the DMCA take-down notice have to conduct a fair-use analysis before delivering the notice to the operator’s DMCA agent?
- What constitutes a legally adequate repeat-infringer policy? What’s an “infringer”? How many times is too many? What constitutes an act of infringement? Does the nature of the underlying work matter? And what about fair use?
- Is any generalized knowledge by the operator of infringing activity on the website defeat the protection, or must there be knowledge of specific acts of infringement?*
- Given that there’s no duty to investigate one’s own website for infringing activity, what does it take to achieve “red-flag” knowledge?
- Is the Financial-Benefit Requirement just a codification of vicarious liability**?
- What does it mean to have “the right and ability to control” infringing activity? Does it mean merely have the ability to block uploads (which almost all websites can do), or to stop specific acts of infringement (which would tend to import the Knowledge Requirement into the Financial-Benefit Requirement)?***
- Is the Section 512(c) safe harbor limited to just the act of storing material, or does it extend to acts involving the access of the stored material, or perhaps even further?*
** Vicarious liability is a judge-made form of secondary copyright liability–i.e., you’re not the actual direct infringer, but you should held liable all the same–in which you gain a direct financial benefit from the infringing activity and you had the right and ability to control the infringing activity, even if you did not know the activity was infringing. The classic example is a dance-hall operator who make money when a band plays copyrighted works without authorization. The idea is that the money the dance-hall operator made really belongs to the copyright holders. The language of the Direct-Benefit Requirement is almost exactly the same as the most common formulation of the elements of vicarious copyright liability.
*** I snuck this one in, too. I don’t have an excuse for not bringing it up in the Grooveshark posts.
The Ninth Circuit Tries to Simplify Things
- Generalized knowledge is never enough to defeat the Knowledge Requirement, either under the “actual knowledge” or “red-flag” prongs. For example, the court disregarded an admission by Veoh’s CEO that he knew that people were using Veoh for infringing purposes.
- The Financial-Benefit Requirement is not coextensive with vicarious liability (otherwise the requirement would gut the safe harbor, since most website operators are probably liable for vicarious liability).
- The ability to stop user uploads is not enough to defeat the Financial-Benefit Requirement (i.e., it doesn’t constitute “the right and ability to control [infringing] activity”). Instead, the operator must have the ability to ferret out and stop specific instances of infringement, which essentially imports the Knowledge Requirement into the Financial-Benefit Requirement.
* The rights holders chose to go for a home run and spent most of their appellate brief’s argument pushing for a narrow interpretation of “by reason of storage,” at the expense of attacking the Repeat-Infringer Requirement.
For the Second Circuit, Simplicity Isn’t Necessarily a Virtue
Now the Second Circuit has weighed in with its much-anticipated decision in Viacom v. YouTube. Previously, the trial court in that case had dismissed the rights holders’ lawsuit in its entirety on grounds that YouTube was completely protected by the DMCA safe harbor. On appeal, the Second Circuit issued a complex opinion, affirming parts of the trial court’s dismissal, flat-out reversing others, and punting still others for further development and analysis (which I’m sure has the trial judge doing cartwheels).
Sorting out the Second Circuit’s actual holding is a bit complicated–unless, of course, you were just to skip to the end of the opinion, where the court lays it all out pretty neatly:
To summarize, we hold that:
(1) The District Court correctly held that 17 U.S.C. § 512(c)(1)(A) requires knowledge or awareness of facts or circumstances that indicate specific and identifiable instances of infringement;
(2) However, the June 23, 2010 order granting summary judgment to YouTube is VACATED [i.e., wiped out] because a reasonable jury could conclude that YouTube had knowledge or awareness under § 512(c)(1)(A) at least with respect to a handful of specific clips; the cause is REMANDED [i.e., sent back] for the District Court [i.e., the trial court] to determine whether YouTube had knowledge or awareness of any specific instances of infringement corresponding to the clips-in-suit;
(3) The willful blindness doctrine may be applied, in appropriate circumstances, to demonstrate knowledge or awareness of specific instances of infringement under § 512(c)(1)(A); the cause is REMANDED for the District Court to consider the application of the willful blindness doctrine in the first instance;
(4) The District Court erred by requiring “item-specific” knowledge of infringement in its interpretation of the “right and ability to control” infringing activity under 17 U.S.C. § 512(c)(1)(B), and the judgment is REVERSED insofar as it rests on that erroneous construction of the statute; the cause is REMANDED for further fact-finding by the District Court on the issues of control and financial benefit;
(5) The District Court correctly held that three of the challenged YouTube software functions—replication, playback, and the related videos feature—occur “by reason of the storage at the direction of a user” within the meaning of 17 U.S.C. § 512(c)(1), and the judgment is AFFIRMED insofar as it so held; the cause is REMANDED for further fact-finding regarding a fourth software function, involving the syndication of YouTube videos to third parties.
On remand [i.e., once the trial court gets the case back], the District Court shall allow the parties to brief the following issues, with a view to permitting renewed motions for summary judgment as soon as practicable:
(A) Whether, on the current record, YouTube had knowledge or awareness of any specific infringements (including any clips-in-suit not expressly noted in this opinion);
(B) Whether, on the current record, YouTube willfully blinded itself to specific infringements;
(C) Whether YouTube had the “right and ability to control” infringing activity within the meaning of § 512(c)(1)(B); and
(D) Whether any clips-in-suit were syndicated to a third party and, if so, whether such syndication occurred “by reason of the storage at the direction of the user” within the meaning of § 512(c)(1), so that YouTube may claim the protection of the § 512(c) safe harbor.
We leave to the sound discretion of the District Court the question of whether some additional, guided discovery is appropriate in order to resolve “(C)” (“[w]hether YouTube had ‘the right and ability to control’ infringing activity”), and “(D)” (“[w]hether any clips-in-suit were syndicated to a third party”). As noted above, for purposes of this case, the record with respect to “(A)” (“[w]hether . . . YouTube had knowledge or awareness of any specific infringements”) and “(B)” (“[w]hether . . . YouTube willfully blinded itself to specific infringements”) is now complete.
You might be able to see why I think the muddying of the waters might be temporary. The Second Circuit is punting a lot of issues back to the trial court and is explicitly hoping to get the issues back on summary judgment*.
* It’s a bit unlikely that the it’ll get issue (A) back on summary judgment. That seems too fact-intensive for summary judgment.
OK, that sets us up for diving into the Second Circuit’s reasoning and figuring out where it agrees with, and where it diverges from, the Ninth Circuit’s reasoning in Veoh. We’ll do that next time.
Thanks for reading!