Читать книгу The Artificial Inventor - Luz Sánchez García - Страница 5

The revolution of Artificial Intelligence (AI), in its different manifestations –expert systems, neural networks, intelligent hybrid systems, etc.– brings with it, technological progress in numerous sectors, such as transportation, medicine or business, among others.

Until now, this technology has been used primarily as a tool to streamline tasks performed by humans, providing both accuracy in data analysis and speed in predicting future scenarios. Thus, artificial neural networks have played a leading role in clinical diagnosis and image analysis due to the parallel processing power that allows networks to learn from historical examples and known patterns1. AI techniques such as generative algorithms can also be used to design new products (eg, lighter structures for use in aircraft interiors)2.

Therefore, we begin to talk about the «Artificial Invention Age»3, an era in which collaboration between humans and machines4 becomes a key point, the former being in charge of detecting the problem and making it concrete and the latter of generating, simulating and evaluate possible solutions5. But it is already a certainty that advanced AI systems are capable of achieving innovative and unpredictable results in an independent and autonomous way, without following any human instruction6, and inventions made by intelligent agents can now be noticed7.

Thus, the scientist and computer scientist Stephen Thaler developed the so-called «Creativity Machine» which, according to his own creator, is the true inventor of the subject matter of the patent «Neural network based prototyping system and method»8, which he applied for in 1998 and of which he claimed holder9. Likewise, the computer scientist John Koza created the so-called «Invention Machine» based on a genetic programming system («Genetic Programming») which, independently, has come to develop numerous patentable inventions10, among which is worth noting, the invention «Apparatus for improved general-purpose PID and non-PID controllers», patented in the United States in 200511.

This panorama, made up of both scenarios in which artificial agents cease to be mere instruments at the service of the human being to become the true protagonists of the inventive process, as well as others in which they remain in the background but their collaboration is essential to reach patentable results, raises controversial questions that should be outlined first technically to later suggest the appropriate legal treatment for its special characteristics. Thus, various questions arise: can these intelligent agents be considered inventors? Could their inventions be patented? Who would be the owner of the patent right? Who would evaluate these inventions? All of these issues will lead to a comprehensive transformation of the patent system.

Although, in general, the Law is adapted to the progress of society once implemented, in the area of inventions made by AI it would be advisable for the former to get ahead of the immediate future and equip them with a regulation that considers the numerous ethical, social and economic implications that this phenomenon may entail12.

In our opinion, this challenge should not be faced from the point of view of fear or uncertainty, but rather suggesting a new approach that allows us to propose appropriate legal solutions to a still emerging reality13. For this, it is necessary to review the traditional concepts of Intellectual and Industrial Property –specifically, in the field of Patent Law14– and to propose new tools in those areas where the former are insufficient or just do not find their perfect framing.

In any case, it is essential to be cautious in the regulation that may be proposed, always considering the respect for human rights15. Hence, that this research proposes an approach to the protection of inventions obtained exclusively or with the help of AI, flexible in terms of the inventive process –in the sense of not limiting the capacities of artificial entities and thereby promoting innovation–, but at the same time restrictive regarding the granting of rights and duties to these entities as a consequence of their action («human-control approach»).

The present work, structured in four chapters, highlights the possible legal problems involved in the traditional treatment of inventions made by artificial entities and, consequently, proposes a new legal framework for the protection of this reality. The first chapter begins with an approach to AI, trying to delimit this discipline conceptually and historically and reviewing the categories it encompasses. We observe that robotics is the branch that a priori best suits the sphere that interests us, that is, the one that contemplates entities with the capacity to invent. In this sense, the chapter focuses on the state of its regulation at the global level, fundamentally influencing its civil aspects, since it is the most developed sphere up to now. Subsequently, with the aim of delimiting in greater detail the artificial agents that could affect the inventive process, the second chapter proposes a new concept, that of Artificial Intelligent Agent (AIA), of which it will be necessary to observe its capacity to be the owner of rights and duties and, with it, that of its possible inventor status. This last issue is the subject of development in the third chapter, where we propose the incorporation into our patent system of a new category of inventor –the Artificial Inventor– that allows us to subsume the invention generated by an AIA among patentable inventions. Finally, in the fourth chapter, a patent ownership regime is proposed based on the degree of participation and autonomy of the artificial intelligent agent in achieving the aforementioned invention.

1. RAMESH, A.N./KAMBHAMPATI, C./MONSON, J.R.T./DREW, P.J., «Artificial Intelligence in Medicine», Annals of the Royal College of Surgeons of England, n. 86, 2004, p. 334-338.

2. See https://www.xataka.com/vehiculos/disenan-un-ala-mas-ligera-para-un-boeing-777-imitando-las-de-las-aves. On this and other advances that are expected to be achieved, see also https://www.autodesk.com/customer-stories/airbus (both accessed 15 January 2019).

3. This term is used by MCLAUGHLIN, M., «Computer-Generated Inventions», Social Science Research Network Electronic Journal, January 2018, pp. 1-23, p.8.

4. On the different positions of what can be understood as “machine” as to whether they can think, see SEARLE, J., «Minds, brains, and programs», Behavioral and Brain Sciences, vol. 3, n. 3, 1980, pp. 417-457; NEWELL, A./SIMON, H., «Computer science as empirical inquiry: symbols and search», Communications of the Association for Computing Machinery, vol. 19, n.3, 1976, pp. 113-126.

5. PLOTKIN, R., The genie in the machine: how computer-automated Inventing is revolutionizing law and business, Ed. Standford University Press, California, 2009, pp. 1 onwards.

6. YANISKY RAVID, S./LIU, X., «When artificial intelligence systems produce inventions: an alternative model for Patent Law at the 3A Era», Cardozo Law Review, vol. 39, 2018, pp. 2215-2263, p. 2220.

7. ABBOTT, R., «I think, therefore I invent: creative computers and the future of Patent Law», Boston College Law Review, vol. 57, n. 4, 2016, pp. 1079-1127, p. 1080. The autor calls them «computational inventions».

8. US Patent n. 5.852.815.

9. About this system –Creativity Machine– based on neural networks, see infra, Chapter I.

10. KOZA, J.R., «Human-competitive results produced by genetic programming», Genetic Programming & Evolvable Machines, n. 11, 2010, pp. 251-284, p. 265.

11. US Patent n. 6.847.851. On this statement see ABBOTT, R., «I think, therefore I invent» cit., p. 1087. Also, KEATS, J., «John Koza has built an Invention Machine», Popular Science, 2006 <https://www.popsci.com/scitech/article/2006-04/john-koza-has-built-invention-machine> (accessed 15 January 2019).

12. On these effects of artificial intelligence and, in particular, robotics, as well as on the relevance of regulating on this point, see the General Principles of the European Parliament Resolution of 16 February 2017, with recommendations to the Commission on Civil Law Rules on Robotics (2015/2103 (INL)). In relation to the need to regulate this new phenomenon and the different legislative techniques to be used, see PAGALLO, U., «LegalAIze: tackling the normative challenges of artificial intelligence and robotics through the secondary rules of law», New Technology, Big Data and the Law (Eds. CORRALES, M./FENWICK, M./FORGÓ, N.), Ed. Springer, Singapur, 2017, pp. 281-298, pp. 287 onwards. See also SANTOS GONZÁLEZ, M.J., «Regulación legal de la robótica y la inteligencia artificial: retos de futuro», Revista Jurídica de la Universidad de León, n. 4, 2017, pp. 25-50, pp. 27 onwards.

13. This position, but in relation to technological advances in the sphere of 3D printing, is the one of LITTLE, R.K., «Guns don´t kill people, 3D printing does? Why the technology is a distraction from effective gun controls», Hastings Law Journal, vol. 65, 2014, pp. 1505-1514, p. 1510. The author understands that control measures need to be taken, allowing, therefore, development and innovation.

14. CLIFFORD, R.D., «Intellectual Property in the era of creative computer program: Will the true creator please stand up? », Tulane Law Review, vol. 71, 1999, pp. 1675-1703, p. 1702.

15. In this regard, about the existing connections between AI and human rights, LATONERO, M., Governing Artificial Intelligence: upholding human rights and dignity, Ed. Data & Society, 2018 <https://datasociety.net/wp-content/uploads/2018/10/DataSociety_Governing_Artificial_Intelligence_Upholding_Human_Rights.pdf> (accessed 15 January 2019).