More on BitTorrent and Grokster
Ernest Miller notes that I should address the new trackerless BitTorrent and BitTorrent search created by Cohen. He was right. So, here goes.
As Wired News reported a few weeks back, Bram Cohen and fellow developers released two innovations to BitTorrent: First, a BitTorrent search engine, and second a trackerless version of BitTorrent. These innovations create toughter questions than the original BitTorrent.
First, the search engine is BitTorrent’s first commercial venture, as it accepts advertisements. If you type “star wars” into the search engine, you find links to several torrents, one of which apparently is a pirated version of Revenge of the Sith. (Here’s a link to the search, but don’t expect the results to stay the same.) What I found interesting is that Sith was only one of the results. The rest appear to be for fan films, which Lucas allows (with occassional equivocation) fans to create and distribute. So, most of the results I found were legal. I suspect this relatively clean result means that Cohen et al. are policing the links. (I may have chosen a biased example, since the BitTorrent folks must know that BitTorrent use in the pirating of Sith has made the movie industry very angry.)
It appears that the new BitTorrent search engine would qualify as an “information location tool” under 17 U.S.C. § 512 (d). This section of the Copyright Act provides a safe harbor from liability for search engines and similar services that comply with its provisions. For qualifying services, it sets up a notice and takedown regime. The problem is that section 512(d) is not an absolute shield. To qualify, the service provider must not have “actual knowledge” of infringement or “aware[ness] of facts or circumstances from which infringing activity is apparent.” If BitTorrent.com becomes an open and notorious source for links to infringing torrents, it could face trouble. At the very least, they are likely to spend a lot of time taking down links, as Mark Lemley notes in the Wired article.
What probably will matter most in the long run is the behavior of Cohen et al., rather than how the search engine functions. My hope (and I believe a reasonable reading of Grokster) is that Grokster establishes a principle that people, rather than technologies, do bad things. Looking at the three key “bad behaviors” from Grokster (see my earlier post), we can see a roadmap for “good behavior.”
The first question is whether BitTorrent.com is directing its promotion its service to large numbers of infringers. As detailed in my earlier post, Cohen and his fellow developers have a long and credible history of building and promoting BitTorrent for legal use. They continue to talk the right talk. As Wired reports,
“the company is eager to highlight its utility as a completely lawful program for furthering free speech. That’s the vision that drives the company, says [COO Ashwin] Navin — now anyone can publish their own movies, music or software, because BitTorrent all but eliminates expensive bandwidth costs.”
Cohen et al. will need to be very careful not to promote illegal filesharing, but they have done a good job so far. They have control over the message they present–they need to maintain a steady, moderate line of promoting legal uses. Greedy talk of how people can download lots of music or revolutionary talk about overthrowing the music industry would only get them in trouble.
BitTorrent.com can also control how it looks with respect to the third “bad fact” from Grokster–will its business model depend on attracting lots of downloaders of illegal material? In many respects, the new search engine appears to be a lot like Google (in substance and appearance). It helps people find stuff; the stuff in this case happens to be torrents. The question is what is the affect if some (or most) of those torrents happen to be illegal? If BitTorrent keeps its search pages clean enough to comply with the Section 512 safe harbor described above, then it likely is fine. If it does not, then it likely will be known as a vast collection of infringing links, and thus its business model will appear to be based on inducing infringement.
The second of the “bad behaviors” in Grokster is more troubling: The inference of bad intent from the failure to create filtering mechanisms. As Susan Crawford, and others have pointed out, footnote 12 of the opinion seems to discourage the creation of duty to filter: “Of course, in the absence of other evidence of intent, a court would
be unable to find contributory infringement liability merely based on a failure to take affirmative steps to prevent infringement, if the device otherwise was capable of substantial noninfringing uses. Such a holding would tread too close to the Sony safe harbor.” If that understanding holds, then the marketing, promotion, and business model will be the key factors in avoiding inducement.
The new trackerless version of BitTorrent presents similar considerations. To put it simply, this innovation makes it easier for people to set up torrents on their own web sites and blogs. Torrents thus may become even more decentralized and harder to find. Arguably, as Ernest Miller points out, this is a design decision that facilitates infringement, so a court might infer intent to induce. As I noted above, footnote 12 of the decision likely makes the marketing and business model aspects more important than the design aspects. As I have said, Cohen et al. have always placed the right emphasis on legal sharing and continue to do so, describing the trackerless version as embodying “our hope that BitTorrent will enable more independent web publishing.” As Ernest Miller puts it, “Bram Cohen must remain purer than Caesar’s wife.”
Ultimately, none of this will stop BitTorrent’s developers or any other developer from getting sued. As I told my clients in practice, if somebody wants to sue you, they will.
UPDATE: See here for further blogging on “shocking” revelations about BitTorrent.