Does IP legalese anywhere stipulate that only human-authored works are copyrightable or patentable? Not according to John Koza, an early pioneer of genetic programming, who is banking on software overtaking wetware in the knowledge-production race. Koza cites the following language to claim that under US law "inventions created by automated means are patentable": “Patentability shall not be negatived by the manner in which the invention was made.” (35 United States Code 103a). Armed with this gift horse of legal ambiguity, Koza filed for patent protection in 2002 on two inventions generated from genetic algorithms, i.e. computer algorithms modeled on Darwinian principles of natural selection.
Koza, it turns out, is a Stanford University professor. I wonder if he's ever crossed the campus quadrangle to shoot the breeze with Lawrence Lessig? Now *there's* a conversation I'd like to overhear . . .
Posted by karik at December 5, 2004 9:33 PM | TrackBackI swear. Pretty soon, I'm going to have to pay a license fee to breathe.
I actually don't have a HUGE problem with this considering that, at least from what I've read, the code used was Koza's own. Well, OK, having typed this, I can see where some real issues might arrive.
But I think that this opens the door for some very directed debates about the importance of interface, code and algorithm in the construction of intellectual property. If we deny Koza's claim, then who's to say that any work produced from Word, Photoshop, Dreamweaver (and more enabling programs such as Flash) aren't also in some way denied in the context of user production?
I know this is a bit extreme but we all know that, say, the interface of Photoshop hides some extremely complicated mathematics that allow for the image manipulation. The math isn't ours-- what we do with it cum interface in some ways is.
Talking in circles here.... I guess that to deny Koza's claim is to keep us free but this problem really exposes some blind spots. Can we deny that the product of self-produced code isn't truly ours? Where does this stop? The hardware itself? Dr. Lessig? Are you there?
Very interesting, Marc--thanks for that. I'm still mulling the analogy between genetic programming and something like Photoshop, trying to decide how well it holds up (do Photoshop or MS Word "automate" the creative process the way genetic programming does?) Since your argument hinges on code, let me pose a slightly different question: do my parents have copyrights in whatever I produce by virtue of "authoring" my genetic code? Yes, I'm shifting the discourse onto a biological plane and literalizing what was for Koza metaphoric, but the hypothetical is continuous with the initial set of issues. (What's the line in Ghost in the Shell? "DNA is a program designed to preserve itself.") Just to raise the stakes, let me point out that apparently DNA *is* copyrightable. On those grounds, aren't I just a derivative work of my parents?