Hello!!

Hi, Robert Plotkin!! I knew that you are linking to my blog and I'm so glad with that. :D I'm sorry for not link, so far, your blog from mine.

Very interesting your ideas about Evolutionary Computation and patents. Once, I was talking with a friend, he is a Lawyer, about a thing that sounds, at a first glance, so "unlikely": Evolutionary Computer Virus. It would be a kind of computer virus that could modify itself through the features of Evolutionary Computation. I think that it is a very interesting subject and, later, that conversation drove us (me and my friend) to another interesting subject: Evolutionary Algorithms applied in indirect Computer Virus Design. Let's imagine that a person makes an Evolutionary Algorithm(EA) to make other EA's, in the set of new EA's produced there are some that works like Computer Viruses do, and those viruses are spread by themselves inside the WWW and, how they are EA's too, they produce new variants of themselves. So, in this situation, who would be the guilty: the person or the EA ? Let's consider that the goal of the person, who made the first EA, was not to produce Computer Virus, but only to produce new EA's. So, How the Lawyer would proceed to find a guilty ? And the Computer Viruses' variants, that work like their parents, like a Computer Virus, who would be the responsible for them ? I think that, at this point, the Law leaves us in a NaN. :D

Até Mais!!

Nosophorus(a.k.a Marcelo de Brito)

Your question is one example of the way in which computer automation is further separating programmers from the results produced by their programs. Whether someone could be found legally responsible in the kind of case you describe would depend on the legal claim that is brought against them. But if, for example, the programmer is sued for negligence, I think it would be hard to prove that the harm caused by the program was foreseeable to the programmer. Therefore, the programmer would have a good defense. But this is likely to play out in different ways under different circumstances. I discussed some of these issues in the "Fighting Keywords" paper that I wrote; you can find this under the "Publications" page on this web site.