Interesting not just because of the intersection of class actions and copyright, but because neither side apparently had the incentive to clarify matters with respect to statutory damages, thus enabling some sloppy language, though nothing that ought to make a difference to this case.
The court began by suggesting that this case was a “Frankenstein monster posing as a class action” (citing Eisen v. Carlisle & Jacquelin, 391 F.2d 555, 572 (2d Cir.1968) (Judge Lumbard, dissenting from remand)).
The putative class consists of every person and entity in the world who own infringed copyrighted works, who have or will register them with the U.S. Copyright Office as required, whose works fall into either of two categories: they were the subject of prior infringement which was blocked by YouTube after notice, but suffered additional infringement through subsequent uploads (the “repeat infringement class”), or are musical compositions which defendants tracked, monetized or identified and allowed to be used without proper authorization (the “music publisher class”).
The plaintiffs claimed that there were at least thousands of members in the repeat infringement class, and hundreds in the music publisher class. (Let me pause here and note that everything about statutory damages that follows would make much more sense if the class definitions had involved works registered before the infringement began or within three months of publication; it is hard for me to imagine that this would’ve changed the size of the class very much except perhaps as to foreign works, since commercially significant US works tend to be timely registered and foreign owners generally either register in timely fashion or they don't, since late registration is pointless for them. Of course plaintiffs probably want a lot of non-US works in the class—but overall then, it probably would’ve been to their benefit to point out that many non-US works in the class wouldn’t be eligible for statutory damages. The idea of eligibility for statutory damages might seem to make damages calculations easier and thus class treatment more acceptable, but here I don’t think that adds much, given that individualized damages calculations are often acceptable if a class is otherwise valid.)
Ahem. Okay: plaintiffs didn’t explain how the worldwide members of the class would be identified, how they’d prove ownership, or how they’d prove YouTube’s awareness of infringement, which is important given YouTube’s general DMCA protection. Given the number of clips on YouTube, “[t]he suggestion that a class action of these dimensions can be managed with judicial resourcefulness is flattering, but unrealistic.” (Of course, neither will individual actions get much of anywhere, even if the individual is Viacom.)
In general, “copyright claims are poor candidates for class-action treatment.” Though of course the elements of infringement/secondary liability are the same across works, the issues “must be resolved upon facts which are particular to that single claim of infringement, and separate from all the other claims.” Aggregation therefore doesn’t simplify dispute resolution. Plus, the economic need to aggregate that often justifies class actions isn’t as significant because “the availability of statutory damages is designed to give litigation value to each individual case.”
Going through the class certification factors in Rule 23(a), the class wasn’t just numerous, but so numerous that joinder was impracticable. “While one can often phrase questions of law or fact in ways that make them ‘common’ to the class, in this case one can do that only at a level of generality which is useless in practical application.” Common questions there might be, but common answers couldn’t be found because of dissimilarities within the proposed class: in particular, whether the copyright owner’s notice was sufficient to permit YouTube to identify and take down the infringing material, along with other defenses like fair use. Other issues: validity and ownership of the copyright; whether the party asserting a claim is authorized to do so; amount of damages; substantial similarity and fair use. These were all too individualized. “One piece of music is unlike another, and is untouched by what infringes the other.” Under Dukes, “the few truly common issues, which largely pertain to the defendants' conduct, do not predominate over individual issues.”
This also meant there was no typicality. “By their very nature, copyrightable works of art are each unique, and what infringes one work will probably have no effect upon another.” (Even exact wholesale copying? Those are special snowflakes!) Each claim had individual issues of ownership, infringement, fair use, and damages.
Nor was Rule 23(b) satisfied. A class action would be inferior because it would “compress into one mammoth proceeding a universe of individual claims, each with its particular facts, issues and (in many cases foreign) law, much better handled in separate cases where each can receive individual attention.” The fear of expense was mitigated by statutory damages, and in any event “the unique nature of each work and of its infringement cannot be obliterated by its inclusion in a sea of other claims, and the defendants are entitled to contest each of them.” As for the individual questions of ownership, assignment, waiver, and fair use (or fair dealing, I presume), they’d be “better handled in the jurisdictions (often foreign) in which they arise, rather than thousands of miles away.”
Nor did the proposed subclasses help any. The first was the repeat infringement class, for works that are or would be registered or didn’t need registration (i.e., foreign works), whose owners submitted takedown notices but suffered subsequent uploads that could’ve been blocked by screening tools. The second was the music publisher class, for those who owned/controlled musical works that were tracked, monetized, or identified by YouTube, “including because YouTube identified a sound recording of the composition using its text-based or audio-fingerprinting screening tools,” whose works weren’t authorized.
As for the repeat infringement class, YouTube just wasn’t obligated by §512 to screen for repeat uploads; plaintiffs were bound by the Viacom holding to that effect, as they were parties to that appeal. That left “little or nothing” to the repeat infringement class, who at most would still need to litigate their individual issues. Plaintiffs’ implication that compliance with a takedown notice was a tacit concession of infringement ignored §512, which made clear that a takedown was no such thing and contemplated counternotification.
And for the music publisher class, even if §512 were overcome, there’d still be all those issues of ownership—“under the applicable foreign law in the instances of foreign plaintiffs”—lack of authorization, fair use, and nature and amount of damages, depending at least in part on registration timing. “Plaintiffs say that the burden is much simpler for this class than in the usual case, because of the special business characteristics of the class definition, and that may be so. But the showing must still be made, and one plaintiff's will be different from another's.” So there was no typical claim that could provide common answers, only a “diverse and unmanageable aggregation of individual claims, better dealt with separately.”
Plaintiffs also sought certification of two issues: whether defendants had the right and ability to control infringing activity on their website and received a direct financial benefit attributable to that activity, and whether defendants' unilateral syndication of clips to third parties was “by reason of the storage at the direction of a user.” These were dealt with in Viacom, and not to plaintiffs’ advantage.