Taking Freenet Seriously: A Response to Picker on Peer-to-Peer

Randy raises a fascinating question below about the appropriate uses for Peer-to-Peer technologies, and he and Tim Wu have begun an interesting dialogue in the comments. Let me suggest an alternative answer to Tim’s.

Peer-to-Peer technologies have substantial utility in those circumstances where anonymity or decentralization are desirable. So, as Ian Clarke has long argued with respect to Freenet, peer-to-peer can be an effective mechanism for enabling free political speech in those parts of the world that have repressive governments. It is relatively easy for a repressive government to shut down one or a dozen central servers, but virtually impossible for them to shut down all content-hosting peers, unless they’re willing to turn off Internet access altogether. Similarly, with central severs, it is much easier to compile a list of the Internet addresses belonging to content downloaders, but much harder to do this effectively when the distribution channels are peer-to-peer. There is enormous potential for these kinds of technologies to promote freedom and democracy in authoritarian regimes and robust, uninhibited debate in freer societies where legal liability concerns and social norms constrain discourse unduly.

At the same time, anonymity and decentralization have substantial downsides in the speech context. Darknets facilitate child pornography distribution. Peer-to-peer allows privacy-invading MPEG files to spread across the globe in hours, well before any court can intervene with injunctive relief. And anonymity and decentralization on P2P can contribute to the rapid spread of computer viruses, thwarting efforts to control viruses through the imposition of legal liability on content providers and disseminators. 

So we have an environment in which P2P creates substantial speech-related benefits and speech-related harms. In these settings, we can resolve this issue in one of two ways: Compare the magnitude of the benefits and harms (This is what the Ginsburg concurrence in Grokster seemed to want to do – More people want to use Grokster to obtain porn than political theory); or give the benefit of the doubt to the “more speech is better” philosophy, and tolerate many less savory uses of P2P for the benefit of the occasional blessed use (This seems closer to Breyer’s view of Sony in Grokster.). The former is how the law usually handles economic policy questions, and the latter, I think, is how it handles many free speech questions. So doesn’t this all boil down to the question of “What is Peer to Peer use: Economic conduct or speech?”  

Comments

[anon]

Interesting. There are also human rights groups using p2p to help document incidents of rights violations - decentralization comes in handy when the gestapo's at the door.

[Bruce]

So, putting this all together, Randy wants to know when peer-to-peer networks are good for non-infringing mass distribution of content. Tim Wu points out that we should consider not just uses now, but uses that may arise in the future. Lior in the post above proposes that peer-to-peer networks are particularly useful "when anonymity or decentralization are desirable."

Limiting ourselves to non-infringing uses, when is anonymity or decentralization desirable? Government censorship is not really an issue in the U.S. Lack of access to high-bandwidth server space is, for small producers of content, but those producers are similarly unlikely to produce anything that both (a) has a large file size, and (b) is in such hot demand that cheap server space will be inadequate. While occasionally an amateur work catches on (e.g., the "All Your Base Are Belong to Us" remix, http://www.planettribes.com/allyourbase/AYB2.swf , or the most recent instance of a server crash due to "Slashdotting"), use of services such as BitTorrent etc. for distribution of legitimate, widely demanded works is probably stunted by the inefficiencies of decentralization, the cost of producing such works, and the lack of a reliable payment mechanism.

Keeping Tim's warning about T-sub1 in mind, I can think of at least two things that might change that might make P2P viable for some legitimate content. One is the development of a payment mechanism associated with P2P distribution -- perhaps including an encrypted file wrapper. Second is decentralization of high-quality content production. While it is currently possible to, say, make a feature film on a budget of a few thousand dollars, it's only rarely something people want to watch. As video editing and effects technology continues to improve and become more affordable, however, it may become possible to cheaply and easily produce a quality audiovisual work that would have mass appeal. But both of those developments are probably far off in the future. And by that time, the cost of bandwidth will likely have decreased dramatically, well below the cost of using a decentralized distribution system. (See, e.g., this "Trends" paper from the MPAA: http://gullfoss2.fcc.gov/prod/ecfs/retrieve.cgi?native_or_pdf=pdf&id_document=6514084364 .)

[Long]

" use of services such as BitTorrent etc. for distribution of legitimate, widely demanded works is probably stunted by the inefficiencies of decentralization, the cost of producing such works, and the lack of a reliable payment mechanism."

I think this is incorrect. Already you have major software publishers (for example Blizzard in their wildly popular World of Warcraft game) which distribute updates via bittorrent. While P2P in it's current format might not allow a mechanism for monetary exchange for legitimate works, there are plenty of works which are legitimate, widely demanded, but require no explicit payment.

As an alternate business model, perhaps business could follow the current "shareware" model of allowing downloads, but only unlocking features if they pay. Microsoft could distribute copies of Office 2026 via P2P, and perhaps you can view documents for free, but you have to pay to unlock the ability to create your own documents.

Also, as the cost of broadband goes down, it's a pretty good bet that computing power will go up as will the size of files that people will want distributed. 10 years ago, the idea of transfering a 4GB file would've been ludicrious. But then again at the time, we didn't have 4GB files to move around, so who cares? In the future, perhaps bandwith will be cheap enough to host larger files, but I'd be willing to bet that the content will be larger.

[Dave]

I like the connection between P2P and free speech, but I’m not sure that the speech-enabling function of P2P is limited to repressed societies. For one thing, individuals could distribute their speech using a centralized server that’s outside the jurisdictional reach of the repressive regime. Also, the speech-distributive effects of P2P can enable communication even outside totalitarian societies. True that the U.S. government doesn’t engage in repression, but it’s possible that server operators may fear contributory liability that renders them less willing to host files that contain outre social criticism, particularly when aimed at large corporations (e.g., Microsoft). P2P thus has the potential to enable marginal speech even in democratic societies (though this kind of speech could take place through means other than file-sharing, such as email, VOIP, or IM).

As for the interesting question whether P2P is speech or economic conduct, it seems to me that P2P itself is neither of these things, but rather a method of enabling both communication and commerce. This doesn’t mean that free speech concerns aren’t relevant; it means that it might make more sense to think about regulation of P2P in terms of the way courts have approached regulation of other venues that operate as spaces for the exchange of ideas and goods (public forums, libraries, bookstores).

Finally, a small footnote to Bruce’s point that it will be a long time before high-quality entertainment will be made cheaply and transmitted over the internet. One of the breakout hits at Cannes in 2004 was Tarnation, a movie made using a home computer for a cost of $218. Artists such as Negativland (www.negativland.com) have also achieved some degree of notoriety for cheaply made features that are available only as torrent files (and at no cost). These features are, of course, the exceptions rather than the rule.