A Copyfighter’s Musings

Frequent readers know my fondness for discussing third party liability rules, including possible alternatives to the Sony/Napster/Grokster line.  I’m presently outlining a short paper discussing, with a focus on P2P, why the basic insight of that line – remedies tailored to a system’s current architecture, with expansive protection for system’s lacking control – is needed now more than ever and is superior to a negligence rule.  In previous posts, I’ve laid much of the framework for this piece, but I’ve missed one point.

The plaintiffs in Grokster suggest that the P2P providers should have used an alternative design to prevent infringement – namely, they could have used filters.  Let’s assume that filtering could be done without impeding legtimate uses – that is, the fair uses we currently associate with P2P remain possible. (Footnote: Probably, this would mean filtering out, rather than filtering in, copyrights works.) Let’s also ignore the problems I previously noted with the scenario.  Given that filtering works without overtly harming legit uses, why should we worry about such a rule?

To see the potential problems, we must first recognize one essential fact: no filtering list will be comprehensive.  New copyrighted works are created every second.  In turn, a P2P provider like Streamcast can’t simply include one filters list within the software – it must be able to update this list for eternity and disseminate it to all users.  In other words, Streamcast needs to create some form of centralized control.

This will have three related consequences:

1.  The Internet, generally speaking, does not discriminate against particular design choices.  You can make something client-server, centralized P2P, decentralized P2P – the Net doesn’t pick a winner.  Instead, we let people innovate in a variety of ways and reap all the benefits.

This liability rule does pick a winner or, at least, a loser – decentralized P2P designs – and thus chokes off innovation.  You might say that this rule is not that different from telling Napster 1.0 to filter its central index.  But filtering in Napster 1.0 doesn’t dictate that the centralized P2P model cannot exist.  Rather, it attempts to come up with a solution within that particular technological design.  In turn, the variety of potential designs is presevered.

2.  Not only do we end up with less innovations, but we’ll also have less innovators.  One of the benefits of a decentralized design is a reduction in creator’s infrastructure costs.  That is, if you’re an amateur programmer, you can release a P2P program into the wild without having to maintain a central server or even a website after it’s been downloaded enough.  You can come up with a general purpose tool and let others figure out how it can be used best.

This is precisely what happened with Gnutella.  Justin Frankel and Tom Pepper came up with it themselves, released it, and that was it for their involvement.  Gnutella is mostly associated with Morpheus today, but it has also been implemented as a Web search engine called InfraSearch.

3.  With less innovators and innovations, we will inevitably impact legitimate uses.  Because we don’t know what those uses might be, we won’t know what we’ve lost.  Furthermore, if we normally think of the Net as a tool of disintermediation, then this rule would mean re-intermediation.  We would be forcing people to connect to some central source in order to distribute legitimate speech.

One might respond that this centralization would have little impact.  People can use Napster to distribute legitimate speech in the same way they can use Morpheus.  There are two potential problems. First, we’ve contracted the number of channels that someone can speak through – only those who can maintain the necessary centralization can control the channels.  Second, with centralization, the system becomes less reliable in certain ways.  In the most innocent case, I’m talking about a Windows server error or a brown out knocking out a server and thus someone’s legitimate use.  In the worst case, I’m thinking of someone attacking the server purposely to disable some controversial speech – someone distributing political documents, for instance.  That sort of fear has influenced the design of Freenet.

As you can see, we end up with many of the same problems you see with other tech mandates.  It wouldn’t change the actual design of the Internet but it would practically affect the brilliance of end-to-end.  Paraphrasing Professor Lessig, it’d be the “Internet,” but not in any sense the Internet that we know today.

When I’ve discussed this argument with others, I’ve often been asked, “So, what do you think we should do about libel and obscenity and everything else on the Internet? Just let it and piracy run rampant with no way of stopping it?” My first reaction is to step back and note that, sure, the values and benefits of the Internet are often in conflict with other social values.  But do we want copyright alone determining how that conflict is settled?  Are we willing to let copyright make design choices for the Internet’s future?

So: questions? comments? criticisms? Like I said, I’m planning a paper, so I’d love to hear counterarguments.

Filed by Derek Slater at 6:18 pm under General news
1 Comment