The Richter Scale^®

I’ve been catching up on my magazine and journal reading and recently finished the February 2005 issue of MIT’s Technology Review, which has a fascinating article by Sam Williams entitled “Unnatural Selection”.

The article explores using machines programmed with genetic algorithms to create new technology. An example given is NASA having a computer design an antennae for a trio of miniature satellites which feature both wide beam and wide bandwidth. The resulting antennae looks like something of a corkscrew, but is better than anything designed purely by humans (at least so the article implies).

What I found particularly interesting however was in the closing paragraphs, where Stanford University’s John Koza, professor of biomedical informatics, has developed a genetic program to design, among other things, electronic circuits. The paragraphs in particular which drew my interest are:

By the time Koza’s group tested the fourth or fifth versions of their program, however, something even more surprising began to happen: the program kicked out circuit designs unpublished anywhere in the patent literature. Two of these designs—a pair of controller circuits that regu­late feedback—were so original that Koza and his colleagues have taken out patents on them.

As proud as he is of his software, Koza isn’t about to assign responsibility for the new designs to the program itself. The patents credit Keane, Koza, and Streeter, in that order. But there are a few new pseudophilosophical conundrums lurking here: If something is invented with no human near, is it really an invention? Who is the inventor? And if the invention actually works, does it matter if we don’t understand how?

I find this a rather troubling commentary. Section 102(f) of U.S.C. 35, the law governing patents, clearly states:

35 U.S.C. 102 Conditions for patentability; novelty and loss of right to patent.
A person shall be entitled to a patent unless —…
(f) he did not himself invent the subject matter sought to be patented, or...

If it was a machine which invented the new circuit design, then arguably, no person did. Sure, a person programmed the machine at some point, but genetic programming includes by definition evolution of the program. So, arguably Keane, Koza, and Streeter should not be able to file for a patent on something they themselves did not invent. Or should they? In the plant world people are claiming they invented (discovered?) naturally occurring compounds. Perhaps the belief is that as artificial intelligences have no status under U.S.C. 35, then it’s perfectly fine for a human to take credit for discovering the results a machine has produced.

I would love the insight of some others in the patent field on this subject.