On the Patentability of Computer-Generated Inventions

So, so, so very far behind. Its even creeped onto the blog, and in particular into our serialization of The Genie In the Machine. Oh, well, better late than never. Here is the last and final installment of our semi-serialization of Robert Plotkin’s book on automated inventing, and its impact on patent law. Meanwhile, you can find the first three installments here, here and here.

Automated Inventing: Should Computer-Generated Inventions be Patentable?

In my previous entries I have discussed how "artificial invention technology" is being used to invent new products automatically. In this entry I will argue that such computer-generated inventions should be patentable, but that the legal requirements for patentability will need to be recalibrated in light of artificial invention technology.

For an invention to be patentable, it must (among other things) be "nonobvious." Imagine, for example, that you design a new pencil. Assume that all previous pencils have been constructed from pine wood and that your pencil is constructed from oak, but in all other respects is the same as an existing pencil. Should you be entitled to a patent on your pencil? It satisfies patent law's "novelty" requirement because it differs in some way from previously-existing pencils. Yet it most likely does not satisfy patent law's "nonobviousness" requirement. A product is considered "obvious," and therefore not patentable, if the product design would have been obvious to a "person having ordinary skill in the art" of the product at the time the product was designed.

Patent law's "person having ordinary skill in the art" (PHOSITA), like tort law's "reasonable person," is a legal fiction, intended to represent the current level of skill of people currently practicing in a particular technological field (art). If a court were to determine whether your oak pencil is obvious, it would first ask, "what was the ordinary level of skill of pencil designers at the time you designed your pencil"? The court would answer this question by looking at factors such as the educational degrees typically held by pencil designers and the number of years of experience they have. A typical statement by a court is that "pencil designers of ordinary skill as of January 1, 2009 had a Bachelor's Degree in Mechanical Engineering and 6 years of work experience designing pencils." Based on this finding, the court attempts to determine whether such a hypothetical person would have found it obvious, on January 1, 2009, to construct pencils out of oak rather than pine.

If all pencil designers were to go back to school and obtain Ph.D.'s in Mechanical Engineering tomorrow, courts would recognize the resulting increased skill of the "person having ordinary skill in the art" in subsequent patent cases. The basic effect would be to raise the bar for nonobviousness, thereby making it more difficult for people to obtain patents on pencils.

Creation and adoption of artificial invention technology by pencil designers could have a similar real-world effect on the ability of pencil designers to create new pencils. One person who I interviewed for The Genie in the Machine said that he considers an engineer with a Bachelor's Degree, but equipped with artificial invention technology, to be as effective at solving problems as an unaided engineer with a Ph.D. If pencil designers worldwide were to adopt, and become skilled at using, artificial invention technology, the law of nonobviousness should take this increase in "effective inventive skill" into account. As a result, it should generally become more difficult to obtain patents on new pencils. Intuitively, someone should not be able to obtain a patent on a pencil that could have been created by any pencil designer merely by applying ordinary skill to widely-available invention automation technology.

Yet it is at best unclear whether this is how the nonobviousness standard will adapt to artificial invention technology. The caselaw on nonobviousness pays very little attention to the technologies that inventors use to assist them in the inventive process, focusing instead primarily on inventors' education and training. Even the U.S. Supreme Court's most recent pronouncement on obviousness in KSR v. Teleflex, 550 U.S. 398 (2007), declined to address this issue as directly as it could have. As a result, we are left with incomplete guidance regarding whether the level of skill of those having ordinary skill in an art in a particular patent case will be interpreted in light of the skill that those in the art have at using extant invention automation technology to solve problems. If invention automation technology is not taken into account when calibrating the level of skill of PHOSITA in particular cases, we run the risk of allowing inventions which could be produced using only ordinary skill to be patented. This would run counter to the very purpose of the nonobviousness requirement and of patent law itself.

As with the other topics addressed in my previous blog entry, I discuss the relationship between nonobviousness and invention automation technology in much more detail in The Genie in the Machine. I hope you have enjoyed this overview of the promise of invention automation technology, and of the challenges that such technology rises for patent law. I look forward to your comments and questions and to an ongoing dialogue.