“DVR as a Service” Isn’t Copyright Infringement–Cartoon Network v. CSC Holdings
By Eric Goldman
The Cartoon Network LP v. CSC Holdings, Inc., No. 07-1480-cv(L) & 07-1511-cv(CON) (2d Cir. Aug. 4, 2008)
The Second Circuit has issued an interesting and potentially important ruling that Cablevision’s DVR as a service does not infringe copyright law. This ruling reverses the district court’s summary judgment for the plaintiff and opens the way for Cablevision to roll out its DVR service offering in the Second Circuit.
The good news is that the opinion eliminates the odd regulatory distinctions between DVRs as a device and DVR as a service. The bad news is that to reach this conclusion, the Second Circuit has to override a lot of adverse precedent, and I’m not sure that other circuits will find this panel’s arguments entirely convincing. As a result, it will be interesting to see if Cablevision interprets this opinion as a greenlight for a national rollout.
Thus, while the opinion is good news for DVR service offerings, the opinion leaves open a lot of questions that will have to be answered in the future. I think it’s safe to say that this opinion is hardly the last stop in our journey.
Buffering Isn’t Infringement
Cablevision’s DVR service splits a broadcast feed into two streams, including a “buffer” copy that goes to a router where it is stored for no more than 1.2 seconds as the router looks to see if any consumers have asked for the program to be recorded for them. If yes, the data goes into their private storage areas; if no, the stream is discarded. The court holds that this buffer copy isn’t fixed because it’s not embodied “for more than a transitory duration.”
To reach this conclusion, the court has to fight against a lot of precedent, especially the MAI v. Peak holding that a copy into RAM is fixed, even though that copy may be embodied for even less time than the buffer copy at issue here. The court says that MAI v. Peak stands for the proposition that these short-duration RAM copies can be fixed but are not automatically fixed. The court says that in MAI (without citing any actual facts from the MAI case), the software surely was resident in RAM for “at least several minutes” while in this case the copies exist for only 1.2 seconds, and this factual difference explains the different conclusion regarding fixations.
There is a major slippery slope problem with this conclusion. Is 3 seconds fixed? 10 seconds? I could keep going, and the court deftly side-steps this problem. Nevertheless, this holding offers some promise for certain types of web activity. First, this ruling might excuse copies made by scrapers/robots who download copyrighted pages to extract unprotectable information on the page. This case suggests that the copies made to download the page and perhaps to process it are not fixed, at least so long as they are flushed really quickly (1.2 seconds or less would be good). Second, this case seems to provide another defense to the otherwise problematic argument that web browsing is infringement; so long as the user hits the back button (and kills any local cache) really fast, no fixation of the web page. The opinion deliberately limits itself to Cablevision’s system of overwriting the data, so that may limits its overall applicability, but this case creates a new category of copies that are embodied in a medium but are not fixed, and this offers some hope for defendants.
Users, Not Cablevision, Make the Other Copies
Even if the buffer copies aren’t fixed (and therefore cannot support an infringement claim), Cablevision still stores a copy of the broadcasted works in its storage area, where users can download the programs. There’s no fixation problem with these, so plaintiffs challenge these copies as both impermissible copies and public performances. The court rejects these arguments, concluding that Cablevision is a sufficiently passive entity that the users and not Cablevision are doing the legally significant activity. Thus, Cablevision is at most exposed to contributory liability for these user activities. Because the plaintiffs had waived allegations of contributory infringement, Cablevision gets summary judgment.
To reach this conclusion, the court ignores Cablevision’s active role in setting up its systems and providing ongoing services, including selecting which broadcast channels are DVRable in its system. Instead, the court sees this fact pattern as identical to DVR as a device, where the DVR manufacturer isn’t directly liable for how the DVR is used. This is consistent with the uncited Field v. Google case, but it conflicts with numerous copyright cases where the service provider’s hosting of files gives the provider more legal responsibility over the system usage than a device maker would have. Similarly, the court distinguishes the coursepack cases on the basis that a human employee of the copyshop presses the “copy” button, because here the system works automatically without manual intervention from Cablevision.
Note, of course, that the court didn’t discuss contributory liability, which also raises the ugly and risky question of whether Cablevision users are directly infringing by using the DVR as a service. I think there is helpful language in the Sony Betamax case about DVRing as a fair use, but I doubt anyone wants to see that battle relitigated.
Similarly, with respect to the argument that the distribution of the files from Cablevision’s storage area is a public performance, the court says that Cablevision isn’t “transmitting” as required by the statute because the user is making the legally significant action.
Further, Cablevision’s delivery of the file isn’t “to the public” as required by the statute. This latter conclusion is totally fine with me as a matter of common sense interpretation of those words, but it runs contrary to numerous messy and analytically questionable precedents regarding the central serving of copyrighted works to private spaces, such as Redd Horne and On Command. The court deftly tries to evade those, but after 2 readings I still can’t figure out what the court said. Maybe the third time will be the charm. I think it has something do with the fact that Cablevision encoded each file delivery to its consumers so that each file delivery could be consumed only by a single playback machine. Let me know if you can figure out what the court was saying here and how it might apply to anyone else. Because the ruling seems to let Cablevision freely broadcast third party content to potentially all of its subscribers without constituting a public performance, I think there may be some exploitable holes here.
One more open question: this opinion makes me wonder if the MP3.com opinion from SDNY is still good law. I’d need to go back through that opinion, but as I recall, a lot turned on the fact that MP3.com tried to act as a proxy for its users. Here, the court treats such proxy activities as passive, and perhaps that analysis would fit the MP3.com facts as well.
I’m excited about this opinion because it eliminates some of the legal anomalies between DVR as a device and DVR as a service. In many situations, DVR as a service will be a better consumer experience, and it is unquestionably better for the environment, so I’m happy that this opinion tries to get copyright law out of the way to enable this result. At the same time, the appellate court set up enough conflicts with other precedent, and sufficiently caveated its opinions to address the narrow facts in Cablevision, that I expect this case resolved nothing definitively. That will have to wait the many cases in our future.
Even so, I remain amused (in a cynical way, not a funny way) that the broadcasters are still fighting against giving consumers what they really want, which is to consume their content at the time and place of the consumer’s own choosing. Eventually, broadcasters are going to have to bite the bullet and post their content onto the Internet for viewers to enjoy at their convenience. There always will be consumers who want to consume the content upon first release, but after that, content that’s unavailable to consumers is just wasting away instead of continuing to make money for the broadcasters.