Читать книгу The Artificial Inventor - Luz Sánchez García - Страница 4

«All men have, by nature, the desire to know»

Aristotle, Metaphysics

The history of humanity is also the history of machines; it is true that they were very basic machines at the beginning –like the wedge of the stone age hand ax–, and highly evolved today –like the android robots–; and machines that have gone from needing the physical force of the human being for its functioning to being able to act autonomously and independently, although perhaps they are no longer simply «machines»

From a socio-economic perspective, the evolution and development of machines and technology in general have caused four industrial revolutions. The first one (1760–1840), was marked by the steam engine, the development of ships and railways, as well as the internal combustion engine. In the second (1870–1914) the new transportation (airplane and automobile) and communication (telephone and radio) systems influenced. The third, called the scientific-technological revolution or of the intelligence (1950–2016), was motivated by the birth of the internet and the development of new forms of energy –renewable ones– and their storage. The fourth revolution –or industry 4.0–, which we have been attending since 2016, is based on advances in robotics, artificial intelligence, nanotechnology, biotechnology, quantum computing, the Internet of Things (Big Data), 3D printing and vehicles, and autonomous ships. But even without finishing to assimilate these new realities, experts are already talking about a fifth industrial –or cognitive– revolution, which they refer to as digital sustainability or automated intelligence, in which computers will be fully capable of improving themselves without human intervention. On the other hand, from a physiological, psychological and sociological perspective, the evolution and development of machines and technology in general have caused a profound transformation in the aptitude and attitude of humans, to the point that we are experiencing today a situation of physical comfort, because machines do the hard work, and also of mental or intellectual laziness, because by letting machines «think» for us we lose cognitive abilities, a phenomenon that has come to be known as «Digital dementia» (SPITZER).

If human ingenuity has made technique its ally on the path to the well-being that machines provide and can provide to men, it must also be their ingenuity that designs the limit framework for the expansion of autonomous or semi-autonomous artificial intelligence independent from human control, in order to prevent the uncontrolled development of this new reality and the risk of man’s dependence on the machine. This is a broad and complex debate in which many scientific disciplines are involved, such as ethics, logic, engineering, robotics, artificial intelligence or law, which is not our task in this Prologue to unravel, but which, in any case, should lead us to conclude that artificial intelligence must be a complement to human capabilities and not a substitute for them. This is how the European Parliament understands it in its Resolution of 16 February 2017 with recommendations to the Commission on Civil Law Rules on Robotics.

But in the concrete stage of the evolution of artificial intelligence systems in which we are for now, and for those non-experts, the information and news related to that reality do not cease to evoke the fictional stories embodied in literature and the movies. Thus, the complete series of Isaac Asimov’s work dedicated to robots (Robot series, 1950-1985); Philip Kindred’s novel Dick Do Androids Dream of Electric Sheep? (1968), who inspired Ridley Scott for Blade Runner; or Daniel H. Wilson’s satirical novel titled How to Survive Robot Uprising: Tips on Defending Yourself Against the Coming Rebellion (2005). And in the cinema, the billboard is made up, along with the aforementioned Blade Runner (1982), and in a non-exhaustive selection, the films Metropolis, by Fritz Lang (1927), Star Wars by George Lucas (1977), The Terminator, by James Cameron (1984), or AI Artificial Intelligence (2001), finally shot by Steven Spielberg, although it started as a Stanley Kubrik project based on the short story Supertoys Last All Summer Long, by Brian Wilson Aldiss (1969).

The fact of bringing up these cultural manifestations, although it may seem frivolous or trivial, is not –or does not seem to be– from the moment when even in the documents of the European Union are cited as principles of robotics those devised by Isaac Asimov in the fantasy of his first work on robots (I, robot, 1950). In short, the imaginative and creative capacity of the human being from the perspective of works of the spirit, musical or plastic has its counterpoint in its inventive and innovative capacity from a scientific- technological perspective as far as the forms of artificial intelligence are concerned. And, what happens is, that what we considered science fiction very recently is already a tangible reality. Perhaps it is, precisely, the speed with which scientific advances are produced and happening constantly and continuously, which prevents us from being fully aware that the environment in which we began our life journey does not coincide with the scenario in which we are currently working and neither with that of a future that is no longer near but imminent. In the continent that is existence, content intermingles the known with the unknown, the tangible with the intangible, the current with the projected, the real with the imagined, in a painting in which the human being coexists unconsciously with the technological advances as if they never had been, as if they had always existed, almost without remembering that there was a previous time not too long ago when it was not.

In this new era, artificial intelligence, arising from the immense capacity of human intelligence, coexists with it. But it does not do so only as an instrument at the service of man and dependent on his programming and control, but with the potential to achieve its own results derived from an autonomous and independent performance with respect to its creator. Thus, for example, the Flow Machines platform –an artificial intelligence from Sony, CSL– was able to create and compose two pop-style songs in 2016 from the 13,000 data related to the ways of writing and reading scores, of styles and of musical compositions stored in the Lead Sheet Data Base. The first one is Daddy´s Car, inspired by «Beattle», whose lyrics, arrangements and production are the work of Benoit Carre, as well as the second one, entitled The Ballad of Mr. Shadow. For its part, the algorithm of the Luxembourg entity Artificial Intelligence Virtual Artist (AIVA), based on deep learning and reinforcement architectures, was able –after reading a large collection of classical music– to detect regularities and compose a piece by itself, being the first artificial composer recognized by a music society (the French Société des auteurs, compositeurs et éditeurs de musique –SACEM–). AIVA offers since 2019 the Music Engine product, capable of generating musical compositions of up to 3 minutes in various styles. The art world also already has works painted by an algorithm created by three French artists –members of the Obvious collective. With data from 15,000 portraits painted between the XIV-XX centuries and a discriminatory factor of human and mechanical works, the system managed to deceive itself and believe that its creation had been painted by a person. The first painting in the collection, Le Comte de Belamy, was purchased by a Parisian collector for € 10,000 in 2018 and auctioned at Christie’s Gallery for more than € 382,000. From the field of industrial property, the DABUS (creativity machine) artificial intelligence system, designed by Stephen Thaler, fed with numerous information and general training, managed to obtain in 2019 two patentable inventions (a food container and an alarm device), although its registration, presented by the Artificial Inventor Project, is having difficulties in the different countries in which it has been requested (United States, United Kingdom, Germany, Israel, China, South Korea and Taiwan) because the inventor is not a human.

We are, therefore, before intelligent «machines» capable of learning even from themselves and that, thanks to the use of algorithms, can process millions of data, analyze them with statistical criteria, and make autonomous decisions, without human intervention. Artificial intelligence has thus given rise to Artificial Intelligent Agents, a new reality that the author of this work and whose prologue is presented here, conceives as Performance Imputation Centers in the final part of the Chapter dedicated to them. The work of approach to the AIAs that is carried out in that part of the work is commendable, treating the divergent questions between human and artificial or mechanical intelligence, and offering the reader an interesting range of legal figures and institutions among which one could be tempted to include the AIAs, but deciding with criteria not to assimilate any of them, given their particular substantivity. This analysis is basic for the construction of the lege ferenda proposal that Luz Sánchez García brings to the table of the legal debate regarding the patentability of inventions developed by Artificial Intelligent Agents and, what comes first, about the attribution of the inventor status to such agents. The notion of Artificial Inventor, already pointed out by the author in her oral (and written) intervention at the European Conference of Intellectual Property Researchers, held in Geneva in June 2019, reaches a commendable degree of development in the third Chapter constituting a life line for those inventions that, if the attribution of the condition of inventor to the agent who would have achieved it, is not accepted, then it would pass into the public domain, taking into account the way in which the patent system is legally designed today.

The issue is controversial. Let us remember that both Intellectual Property and Industrial Property have at their center the human being. And so, it is the individual who has the right to be recognized as an author or as an inventor, a moral condition that fits perfectly into a general system of Law built on the basis of the person, subject of rights and obligations, while being a moral entity. It is not surprising, therefore, the reluctance that, from many points of view and to many sectors, arouses the appearance on the scene of artificial intelligent agents that act and achieve results, among other reasons –not banal–, due to the difficult fit that all this has in the western legal conception of Law as a science created by and for humans (the human being is the center of Law).

It is true that not every new phenomenon necessarily requires new rules specifically and especially dictated to respond to the also new situations that it originates. The concrete legal arrangement of a subject is understood to be elastic and versatile enough to be applicable, correctly interpreted, to the new demands of reality. Now, when this arrangement has at its base, or as a center or axis, a very specific subject, object, circumstance, or fact, the extension of the regulatory norms to a different and even incompatible one, may give rise to unintended consequences by Law or by the socio- economic order, for being harmful. In such a case, the friction of the current rules with the reality in respect to which its application is intended, could erode the first justification of those and their own effectiveness, discouraging the proposed subsumption. And so, it is in the framework of Intellectual and Industrial Property, whose norms, centered on the human author or inventor, reject and expel from their system the works or inventions of the artificial author or inventor. The question, then, is deciding whether to maintain the traditional conception or to accommodate technological evolution and innovation. In this sense, the World Intellectual Property Organization opened in December 2019 a public consultation period (ended in February 2020) on artificial intelligence and intellectual property policies, in which the invitation was to obtain comments on a theme document designed to help define the most pressing issues faced by intellectual property policymakers, as artificial intelligence becomes more important. An analysis of the query and its results is offered by the author –participant in it– on the pages of this work, particularly in the introductory part of the Chapter dedicated to the artificial intelligent agent as a generator of inventions (the artificial inventor).

Based on the ultimate reason for the existence of the patent system, Professor Sánchez García admits both the patentability of the inventions achieved by the AIAs as well as their qualification as inventors, which leads her to demand an adequate treatment of these entities as actors in the inventive process and specific legal protection for its results. In her proposal, the patent system would then be unfolded in attention to the inventor’s human or artificial nature, reserving the current legal regime for the human inventor and his inventions and designing an ad hoc regulation for the artificial one –in its condition of Performance Imputation Center– and their results, with attribution of rights to the creators and users of the artificial agent. According to the author «it would not be possible to speak of a moral right of the AIA (that of being recognized as an inventor), but of a formal requirement for the identification of the true center of imputation of the inventive activity displayed, which would allow locating the actions that gave rise to the achievement of the invention and, ultimately, specify the human subject that supervises it».

This is not the time to reveal the content of the interesting reflections and important proposals that the fourth Chapter of the work contains, for example, on how to detect if the AIA has been a mere instrument of the human inventor or if it has acted autonomously and independently and therefore is the inventor (through the mechanism called Invention’s Conception Contribution Test), obtaining a certificate of artificial inventor; on the patentability requirements of this type of invention and the inevitable entry of artificial intelligence into the expert system in relation to the technical state of the art; on the options in order to attribute the ownership of the inventions generated by the AIAs; on the creation of a registry of these agents in order to allow not only their individual identification and that of their creator, but also that of the person behind it when the invention is generated (the creator or the user); or about the creation of a fund that, with the benefits obtained from the exploitation of the inventions generated by the AIA, can face the consequences –negative– of the inventive process from an economic or patrimonial perspective and to the expenses of conservation, repair or improvement of the AIA, including those of the application, granting and exploitation process of any of its patentable inventions. But despite this, and even at the risk of having incurred in a spoiler with the aforementioned, we could not resist the temptation of arouse in the reader a degree of curiosity sufficient enough to proceed to enter the pages of this work with a critical gaze.

The Artificial Inventor is, in effect, a challenge for the patent system that is bravely and brilliantly addressed by Luz Sánchez García, making this work a benchmark study and a valuable instrument for doctrine, legal and economic operators and, of course, for those responsible for formulating industrial property policies whose interest, for sure, they will appreciate. My most sincere congratulations.

Murcia, 8 March 2020

Rosalía Alfonso Sánchez

Professor of Commercial Law