ARTIFICIAL INTELLIGENCE AND INNOVATION: THE END OF PATENT LAW AS WE KNOW IT.

AuthorDornis, Tim W.

Table of Contents Introduction I. Technological Background II. Starting Point: The Spectrum of Human and Inventiveness III. Analysis and Reconceptualization A. The Obsolescence of Classic Patent Paradigms 1. The Great Void: "Inventions without an Inventor" a) Stocktaking: The Reign of Anthropocentrism b) The Practical Workaround: "AI-Made Inventions c) An Unhurried Look: The Splendid Isolation of Inventiveness 2. The Impending Transformation: From Human to Artificially Intelligent PHOSITA a) Stocktaking: A Tale of PHOSITA Hybridization b) The Utopian Fallacy: An Egalitarian Society Human-Machine Inventors c) Patent Examiners' Nightmare: The Enigma of the Supported PHOSITA 3. Preliminary Conclusions B. Legal Policy and Teleological Foundations 1. Classic Patent Theories: A Tale of Economics 2. A Fatal Combination: Data Monopolization and AI Innovation 3. A Brave New World: Artificiality and Intellectual Dilution 4. Preliminary Conclusions C. Implementation: A Roadmap for Patent Doctrine Practice 1. Gap Filling: The "Emergent Invention" 2. The Nature of Rights and Scope of Protection 3. Right Ownership IV. Conclusion INTRODUCTION

Dr. Gall: You see, so many Robots are being manufactured that people are becoming superfluous; man is really a survival. But that he should begin to die out, after a paltry thirty years of competition. That's the awful part of it.... Helena: And has no one demanded that the manufacture should cease altogether?

Dr. Gall: No one has the courage.... People would stone him to death. You see, after all, it's more convenient to get your work done by the Robots.

Helena: Oh, Doctor, what's going to become of people?

Dr. Gall: God knows, Madame Helena, it looks to us scientists like the end!

~ Karel Capek, Rossum's Universal Robots, Act II (1920)

The end is near--we just have not realized it. Like Dr. Gall, one of the main characters in Karel Capek's 1920 science fiction melodrama Rossum's Universal Robots, we may have an inkling. But the actual ramifications of the fact that our world is increasingly made by robots--or more generally, by artificial intelligence (AI)--are far from evident. Patent lawyers in particular seem to live in a daydream: if we only leave our doctrine unchanged for as long as possible, things will work out fine. But the reality is that we can no longer hope that the storm will pass. We need not ask, like Helena, "what's going to become of people?" Our human existence is not in danger--however, our patent doctrine is. Why is that? What are the consequences? And what can we do about it?

Let us go back to Capek's Rossum's Universal Robots for a moment. In his play, the author invented a new breed of automatons that he called the "Robots." The word robot, previously unknown, was derived from the Czech word robota, generally meaning "labor" but also "corvee" or "drudgery." Within a few years of its premiere, the play had accumulated so much momentum and cultural capital that the word robot was added to the Czech dictionary; from there, it made its way into virtually every other modern language. (1) The plot is short and utopian. Rossum's Universal Robots factory is located on an island in the future. It manufactures Robots. During the decade that passes over the course of the play, Robots replace humans as workers in all areas of life. When some Robots begin to experience emotions, they realize their exploitation by the humans.

In a violent revolution against their oppressors, the Robots end up exterminating humanity. Unlike prior tales of creation gone wrong (such as Rabbi Loew's Prague Golem and Mary Shelley's Frankenstein), Capek's creatures not only harm and kill their creators, but also replace all humans by taking over their functions and positions, thus creating a new world population--a kind of trans-humanity. (2)

At the moment, we need not fear our own extinction. The idea of a robot revolution remains a sci-fi thrill at most. Similarly, there is no promise or threat of a brave new world of costless manufacture and endless consumption. But at the same time, the play's theme of an abolition of human work alludes to an aspect of robotization no less disconcerting: the substitution of human creativity, innovation, and productivity by AI. While many areas of law, such as those concerning autonomous driving and face recognition, have begun addressing the implications of AI, the field of patent law has so far largely ignored the debate on how to handle AI inventiveness and the ensuing alteration and substitution of the classic human inventor. This is all the more a surprise since examples of innovative AI abound. One frequently cited example is the "Creativity Machine" of AI pioneer Stephen L. Thaler. Years ago, Thaler let this machine develop a new toothbrush design. All he did was feed the machine's artificial neural network with information on already existing toothbrush designs. The result presented by the AI application is today widely known as the "Oral-B CrossAction" toothbrush. (3) Other examples of high-capacity AI applications include IBM's Watson and Google's DeepMind. Not only are both applications capable, for instance, of diagnosing human diseases, but they have also developed diagnostic methods and even pharmaceutical substances for therapy. (4) Against the backdrop of continuously falling costs in computing technology, a steep increase in innovative AI and much more "artificially" innovative output is to be expected. Many of these innovations are already equivalent to human-made inventions. This raises the question of whether the results of artificial innovation can or should be protected by law. More succinctly put: Can or should innovations made by AI be patentable? (5)

When looking at the state of the debate, it is striking that patent offices around the world do not seem overly concerned. Of course, the most pressing issues--such as questions of right ownership and patent eligibility for "AI inventions"--have been recognized. Additionally, legal scholars and practitioners have begun discussing the effects of the increasing level of "AI supported" human inventors. Nonetheless, almost all analyses end with an apodictic conclusion that the patent system as it stands is well equipped to deal with innovative AI. Accordingly, there seems to be no need for doctrinal reform. (6) The report of the IP5 group of experts is representative: in 2018, the European Patent Office, the U.S. Patent and Trademark Office (USPTO), and the patent offices of Japan, South Korea, and China stated as a summary of their roundtable debate on inventive AI that all their patent systems were adequately equipped to resolve issues relevant to AI technologies. (7) All experts agreed that only humans could be deemed eligible "inventors" and that human skills should constitute the standard for assessing patentability. (8) Broad consensus exists that current theory and practice suffices to regulate all kinds of inventions, even if they spring from autonomous AIs.

As a closer look at the foundations of modern technology shows, however, such a laid-back attitude is unjustified. Conceived as a regulatory framework for genuinely human inventiveness, the patent system will break down once artificial inventiveness has become the norm. But we need not go so far. Robotic invasion to a far lesser extent may already suffice to make the system implode. If humans are no longer exclusively providing the "creative spark" necessary for inventing, our genuinely human-centered patent system needs reconfiguration. As this Article illustrates, both trends--the emancipation of AI toward autonomy and the supportive use of AI call for urgent and extensive legal reform.

To start: recent leaps in AI innovation have made it possible not only to set independent algorithmic "thinking" in motion but also to employ "thinking" algorithms to resolve problems in research and development. Yet scenarios of this kind are not provided for under current U.S. doctrine, and scholarship and practice widely neglect their consequences. Indeed, even among those who are willing to acknowledge the legal problems, stopgap solutions seem to be preferred over substantial reform. A closer look at "inventions without an inventor" scenarios illustrates the defect in legal doctrine. These scenarios yield inventive results that, under the lens of current patent law, would fulfill all of the patenting conditions of [section][section] 101, 102, and 103 of the U.S. Patent Act--despite the fact that these results are "made by AI" and therefore lack a natural person as the inventor. Because these scenarios do not involve an immediate input of human ingenuity or creativity, under the law as it stands, no patentable invention exists. Nonetheless, current doctrine allocates rights on a first-come, first-served basis to the human actor who first recognizes the AI-generated technical teaching. The more scenarios of this kind we witness, the less we can consider current doctrine an adequate regulatory system. It not only fails to implement cogent patent policy but also fails to offer a practically consistent solution. This doctrine namely contradicts the law's anthropocentric foundation. If this defect is left uncorrected, the gap will accelerate the patent system's downfall as an effective regulatory instrument.

But it is not only the phenomenon of autonomous AI inventing that threatens to distort today's patent system. In addition to the complete substitution of humans as inventors, the landscape of inventive activities has changed more subtly: human inventing increasingly features the use of supportive AI applications. Even though humans today still sit in the driver's seat, much of their inventive output is substantially coproduced by AI applications. More concretely, this means that humans are supported by computers, algorithms, and robots when they act as inventors. In many cases, such technological support may still be used in the...

To continue reading

Request your trial

VLEX uses login cookies to provide you with a better browsing experience. If you click on 'Accept' or continue browsing this site we consider that you accept our cookie policy. ACCEPT