Auto-Prior-Art: Clever, But Probably Won't Work



All Prior Art has drawn attention from Gizmodo and National Public Radio, among others, for its attempt to use math to create new prior art and publicly license it so new inventions are harder to patent. 

The catch is that it’s missing a key patent law ingredient. 

Alexander Reben’s project is designed to break down all published patents and applications into text snippets and then combine the snippets in every possible way. 

The virtually infinite number of combinations will generate over 99 percent dreck. But, like the “infinite monkey theorem” that predicts a monkey hitting typewriter keys at random for an indefinite time will almost surely, eventually type a given text, Reben believes his project will generate at least some combinations that someone, someday might try to patent. 

It’s that someone he aims to stop. 

If things go according to Reben’s plan, a prospective patent holder will fail because an algorithmically-produced combination will be used as prior art to invalidate his invention—either because the Patent and Trademark Office finds it or the first person accused of infringement uses it as a defense. 

The problem is that, for any prior art reference to serve as an invalidating disclosure, it has to be “enabling.” The means it must be described so that a “person of ordinary skill in the field of the invention can practice the subject matter based on the reference, without undue experimentation.” (Sanofi-Synthelabo v. Apotex, Inc., 550 F.3d 1075, 1082 (Fed. Cir. 2008). 

When it comes to All Prior Art’s prior art, the “description” is missing. And the fact that no human being came up with the combination is a pretty good argument against enablement. 

To be fair, Reben’s goal is “to take the obvious and easily automated ideas out-of-play” in the patent arena. Presumably, that means that the PTO or a court would tell a prospective patentee to go back to the drawing board because they should “at least be more creative than a random-combination generator to deserve a patent.” 

Along those lines, Reben’s combinations might prove useful if someone could enable one of them with little-to-no work, triggering the “without undue experimentation” requirement. Then again, if enablement is that trivial, one would think the PTO would reject it outright. 

Most broadly, Reben hopes “to democratize ideas, provide an impetus for change in the antiquated patent system, and to preempt patent trolls” with his project. Laudable goals. But the implementation may fall short.