Rethinking Intellectual Property: Exploring Social Implications of AI-Generated Inventions in Patent Systems

Wednesday, 9 July 2025: 16:00
Location: ASJE032 (Annex of the Faculty of Legal, Economic, and Social Sciences)
Oral Presentation
Benedicto ACOSTA, Universidad de Salamanca, Spain
The integration of artificial intelligence (AI) into various aspects of society has sparked profound questions regarding the ownership and protection of intellectual property rights. This work presents a comprehensive examination of the ethical, legal, and practical considerations surrounding this debate, drawing on relevant literature to provide a nuanced analysis against the patentability of AI inventions. At the heart of the discussion lies the fundamental question of agency: Who should be considered the inventor when AI systems autonomously conceive novel ideas? Traditional patent law predicates inventorship on human creativity and ingenuity, raising doubts about the applicability of existing legal frameworks to AI-generated inventions. Proponents argue that granting patents for AI-generated inventions incentivizes investment in AI research and development, fostering technological progress and economic growth. However, detractors caution against the potential consequences of conferring exclusive rights to entities devoid of consciousness or moral responsibility, which could stifle competition, exacerbate inequality, and hinder access to essential technologies.

Using the framework of the so-called "social contract of patents" I argue: (1) that the social reparation of the social costs of invention through commercial exclusivity is not in the interest of the AI, but only of the owners or of those who were its inventors in the first place, who in any case would suffer a reduced marginal cost; (2) on the other hand, the publicity of the invention through the patent document, which is the result of the exclusivity envisaged by the contractualist defense, is made difficult in this case, given the nature of some artificial intelligences, and in particular of neural networks. This also leads me to examine the more fundamental question, specific to natural law intellectual property defenses, of whether AIs deserve such recognition, which is briefly discussed with the help of two recent court decisions.