Patently-O Bits and Bytes by Juvan Bonni

Recent Headlines in the IP World:

Commentary and Journal Articles:

New Job Postings on Patently-O:

5 thoughts on “Patently-O Bits and Bytes by Juvan Bonni

  1. 3

    OT, but highly relevant to 103, the most commonly argued issue in patent law, and an unusual reversal of an IPR patentable decision: Airbus SAS, Firepass Corporation (Fed. Cir., 8 November 2019). Because the Board incorrectly overlooked a cited as analogous prior art reference on the question of obviousness. Deciding that although the reference was outside the inventor’s “field of endeavor” that reference would have been discovered in the ordinary course of research for solutions for the problem being addressed.

  2. 2

    The lead article here is about alleged AI inventorship issues. That may well be the least difficult AI legal issue to resolve, perhaps as fast as the recent attempted copyright registration by a chimpanzee of his selfie. Machines do not qualify as inventors either.
    Nor under established patent law are persons merely running an experiment or test designed and supervised by the inventor co-inventors.
    One important difference for AI patenting, as opposed to patenting re true pioneer breakthrough inventions, is that AI already has a considerable history of prior publications dating from its first big academic fad era some years ago [when truly massive retrievable data storage, and computations on it, was not yet available]. That may now present prior art issues in some cases, limiting some patents to inventors of novel and unobvious AI algorithms? Any thoughts on that?

  3. 1

    With regards to the AI patenting “issue”, it’s amusing to watch the hand-wringing. Or the pretense of it.

    The house of cards built on a foundation of vapor that is the patenting of logic “on a computer” is already enormous and tilting. The people who make a living off gambling on the ultimate height are totally aware of the absurdities that are baked in already. And they don’t care. They definitely don’t care about piling a few more cards on top now.

    One difficulty when enforcing AI patents is that it is often not clear how an AI system actually works


    “I don’t know why it works” is actually not a problem for the US patent system. The problem for the system is “I don’t know how to describe the difference between this machine and the prior art machines except in terms of a result.” And that problem is “easy” to solve for the titanic legal geniuses who comment here. You solves that problem by waving your hands around and saying “The machine must be different because the result is different from any other machine that ever existed, at least as far as the PTO knows, and therefore I should have the right to sue you for achieving that result using your AI machine which must be the same as mine because it achieves the same result.” Yes, these are very very serious people. They are definitely NOT just entitled glibertarian Silly Con Valley Bros trying to get Rich Whitey’s money (and then everybody else’s money). Nope. They are very serious.

    Just wait until they show up here and try to defend AI patenting. You will be very impressed by their deep thoughts. Trust me on that. The best minds. Only the best.

    1. 1.1

      who can (and should) be entitled to the patent for AI-generated inventions: the AI system or its owner, deviser, trainer, or user?

      Super difficult questions! But why hurry to resolve them? Just grant the patents and then we’ll let the ultra rich resolve them for us.

      This stuff is really complicated, man. Nobody has ever thought about any of this before, in any context.

      Right, Dennis?

Comments are closed.