The Evolving Interface of Intellectual Property and Artificial Intelligence
Jyh-An Lee, The Chinese University of Hong Kong;
Reto Hilty, Max Planck Institute of Innovation and Competition;
Kung-Chung Liu, Singapore Management University
-- The ability of computers to imitate intelligent human behaviour has drawn wide attention in recent years; we humans are increasingly ceding our decision-making power to technological artefacts. With the advancement of data science and computing technologies, artificial intelligence (AI) has become omnipresent in all sectors of society. Face and speech recognition, visual perception, self-driving vehicles, surgical robots and automated recommendations from social media are all well-known examples of how computer systems perform tasks that normally require human intelligence. From manufacturing to healthcare services, AI has already improved previous best practices. Based on large volumes of data, AI can predict more accurately than humanly possible. The overwhelming intellectual power of AI is also exemplified by AlphaGo and AlphaZero, which have taught themselves to beat the best human players of Chess, Go and Shogi.
AI also enables new models of creativity and innovation with its data-driven approach. While human beings have used various instruments and technologies to create and innovate, they themselves have been the main driving force of creativity and innovation. AI puts that into question, raising numerous challenges to the existing intellectual property (IP) regime. Traditionally, the “intellectual” part of “intellectual property” refers to human intellect. However, since machines have become intelligent and are increasingly capable of making creative, innovative choices based on opaque algorithms, the “intellectual” in “intellectual property” turns out to be perplexing. Existing human-centric IP regimes based on promoting incentives and avoiding disincentives may no longer be relevant – and could even be positively detrimental – if AI comes into play. Moreover, AI has sparked new issues in IP law regarding legal subjects, scope, standards of protection, exceptions and relationships between actors.
Our recent book, Artificial Intelligence and Intellectual Property (Oxford University Press, 2021), examines these issues in seven, interconnecting parts. Part A provides the technical, business and economic foundations for the later analysis of IP issues in the AI environment in the following parts of the book. Part B examines emerging substantive patent law and policy issues associated with AI, including foundational patents in AI-related inventions, the patentability of AI inventions and how AI tools raise the standard of the inventive step. This part also illustrates how patent prosecution has evolved from material to textual to digital. Part C probes into two major copyright issues concerning AI’s involvement in creation: the copyrightability of AI-generated works and copyright exceptions for text and data mining (TDM). Parts B and C present various legal and policy concerns in patent law and copyright law, respectively. However, patent law, copyright law and trademark law occasionally share the same conundrum caused by the rapid development of AI technologies.
From Parts D to G, this book covers issues relevant to multiple categories of IP. While AI has enhanced the efficiency of IP administration and enforcement, it has generated new problems yet to be solved. Therefore, Part D explores how AI reshapes IP administration in the public sector and IP enforcement in the private sector. Part E examines copyright and patent protection for AI software, which is qualitatively different from traditional computer programs. While AI is implemented by software, the protection for such software per se has been ignored by the mainstream IP literature. Part F discusses the protection of and access to data, which is the driving force of all AI inventions and applications. It further illustrates how IP law will interact with other fields of law, such as unfair competition law and personal data protection law, on various data-related issues. Part G provides a broader picture of AI and IP, searching for solutions to fundamental inquiries, such as IP and competition policy in the era of AI and whether an AI should be viewed as a legal person.
The book is a result of collaboration between the Chinese University of Hong Kong Faculty of Law, the Applied Research Centre for Intellectual Assets and the Law in Asia (ARCIALA), the School of Law, Singapore Management University, and the Max Planck Institute of Innovation and Competition. It sets out to elucidate the general challenges and opportunities faced by every jurisdiction in the era of AI. We believe all policy and legal analysis should be based on a correct understanding of the technology and the economics of innovation, and an ideal policy should facilitate human sovereignty over machine efficiency. By the same token, a desirable IP policy should enable society to fully grasp the value of new technologies for economic prosperity.