Author: Pavithra Rajendran, Student at SASTRA Deemed to be University, Thanjavur, Tamil Nadu
Robots, Artificial Intelligence, and machine learning systems (hereinafter referred to collectively as “AI” or “AI systems”) can now produce inventions that, if made by humans, would be patentable. This article looks into the patentability of these AI-created inventions. As AI gains traction in the manufacturing, fintech, healthcare, retail, educational, and infrastructure development, the potential for disruption lies not only in the acceptance of AI but also in how it is developed and protected; there is also the more key issue of who owns AI. As the science and business applications of AI for innovation are gaining interest, little consideration appears to have been given to the legal implications, especially in the area of patent law. This article identifies several important issues for patent system participants to address while thinking about AI and the obstacles it will bring to the patent system. Although there is widespread agreement that AI-generated inventions are incompatible with the principle of human inventorship, it is unclear to what degree questions about ‘non-human’ ingenuity can be justified. The most ambiguous aspect is how AI ‘autonomously produces’ inventions, and how ‘AI-generated’ inventions vary from inventions produced with AI’s assistance. The article further outlines how the new paradigm will affect the prevailing patent system in several ways, including conceptually challenging the notions of inventor and invention, accountability, and infringement liability. It contends that the conventional patent law has become obsolete, inapplicable, and meaningless in the context of AI-generated inventions and urges lawmakers to reconsider existing patent law regulating AI systems and replace it with laws better suited to the modern age of advanced, automated, and autonomous AI systems.
Artificial intelligence (AI) is non-sentient machine intelligence that is designed to perform a specific task. AI is intelligence demonstrated by machines, as opposed to natural intelligence demonstrated by humans and other animals. Artificial intelligence has disrupted the entire technological ecosystem and opened previously unheard avenues as it is implemented on a macro level in various industrial sectors. Advanced AI systems are highly capable of producing unpredictable, innovative outcomes on their own, rather than simply following digital commands. These machines are producing inventions that could qualify as patentable inventions. As AI’s role in the invention process continues to grow, it can produce results that ordinarily occur as a consequence of the application of human cognitive processes. The evolving landscape has managed to present challenges for the players and rules that make up the Indian patent system. The debate over ‘AI-generated’ inventions is gaining traction, and it is now on the agenda of international policymakers. Lawmakers and judges must decide how to adapt the existing statutory framework governing patents to keep pace with new technological developments and changing market conditions.
The question of whether the current standard under the Indian Patents Act, 1970 encourages or curtails technological innovations in the electronics/computer space has been debated in India for several decades. An assessment of approximately 80 cases in AY 2019-2020 revealed a clear trend among examiners to treat AI-related inventions as computer programs or algorithms, and thereby deny patents under Section 3(k) of the Act, which lists mathematical, business methods, computer programs per se, and algorithms as non-patentable subject matter.
In the case of Ferid Allani vs Union of India & Ors., the Delhi High Court held that computer-related inventions (CRIs) are not barred from patentability if it has a specific technical purpose. The patent applications in these fields must be examined to determine whether they result in a technical contribution, and if the invention illustrates a technical effect or a technical contribution, it is patentable even if it is based on a computer programme. While this decision, among other things, is a progressive step in the direction of acknowledging creativity exhibited by AI, due regard must be paid to several other fundamental questions that remain unanswered. This decision demonstrates the difficulty of transitioning from a traditional system of the invention to an innovative and as-yet poorly understood model of the invention while continuing to rely on existing patentability rules. which will be addressed in this article. Some of the major patent law issues impacted by AI are described in detail below.
INVENTORSHIP ISSUES FOR AI -GENERATED INVENTIONS
As algorithms become more sophisticated and human interference in the development process reduces, the identity of the creator of an AI-generated invention becomes ambiguous. To apply for a patent, the true and first inventor(s) of an invention must be identified. In such instances, should a human or the machine be named as the inventor or inventors of an AI invention? Under Section 6 of the Indian Patent Act, 1970 only the true and first ‘inventor’ of an invention or an assignee can make an application for a patenting any invention. Further, a patentee is defined under Section 2(1)(p) as the “person” listed on the patent office register as the grantee or owner of the patent. This alleviates any assumption that supports the premise that the legislative intent of Indian Patent Law sought to include inventions or, more specifically, the possibility of inventions being made by anyone other than humans. The idea of a human inventor is embedded in the patent application process as the Indian Patent Act, 1970 is rife with references to human actions and entitlements.
Consideration of AI-generated inventions as patentable and recognition of AI as an inventor would necessitate AI being treated as a legal person. As a legal person, AI would be subject to the rights and obligations that come with that status. The second alternative is to not include any inventors. However, this would necessitate a change in the patent legal framework to allow patents to be granted to AI without naming the inventor. In the aforementioned scenario, adequate measures must be taken to provide incentives to those involved in the creation and maintenance of AI. So that they can continue to develop AI, which generates novel ideas. Hence, the difficulty in tracing an invention to the inventor necessitates legal perusal.
ACCOUNTABILITY AND INFRINGEMENT LIABILITY
Another illustration of indeterminacy in patent law is AI liability in cases where AI is the infringer of patent rights. With the evolving technological landscape, most AIs are now capable of infringing on other patents. The problem of liability raises the question of who should be held responsible for AI’s conduct, whether the AI itself or the AI’s creator and how AI’s liability should be evaluated.
The patent issues for AI-generated inventions must be determined against the backdrop of whether patent rights to AI-generated inventions would advance the goal of patent law or would end up being detrimental to inventions created by humans. Furthermore, the liability that AI must bear in the event of infringement must be determined.
Given the rapid development of AI-related work, companies have begun to rely on this advanced technology to solve complex business problems and develop new applications without the assistance of humans. The existing patent scheme is unsuitable for deciding patentability issues relating to the current generation of AI technology. The time has come for the Indian government to start contemplating about how CRIs pertaining to the next generation of AI (i.e., powerful AI and superintelligence) should be handled in the patent ecosystem. Patent authorities must reconsider their traditional laws and develop new framework rules that include devices, computers, and non-natural persons as inventors. It is ironic that the legal industry, which is inherently a cautious adopter of new technology, finds itself at the cutting edge of technology, dictating rules that will guide the development and control over new technology well before witnessing the transformative impact of such technology on its business practice. It is incumbent on professionals who are involved in the patent process industry to acquire the skills and implement the guidelines required to administer patents in the aftermath of technological advances that are changing the invention process. Appropriate patent policy amendments are now more important than ever. Authorities must act expeditiously to establish a streamlined path; failing which, they risk disincentivizing the creation of path-breaking innovations.
University: SASTRA Deemed to be University, Thanjavur, Tamil Nadu
Course: B.A., LL.B., (Hons)
Academic Year: 4th year
 Liza Vertinsky& Todd M. Rice, Thinking about Thinking Machines: Implications for Machine Inventors For Patent Law, B.U.J SCI.&TECH L. 82 (2002).
 The World Economic Forum (WEF), ‘Artificial Intelligence Collides with Patent Law’ (April 2018) <http://www3.weforum.org/docs/WEF_48540_WP_End_of_Innovation_Protecting_Patent_Law.pdf>.
 Ferid Allani V. Union of India, W.P.(C) 7/2014 & Cm Appl. 40736/2019, Decision Dated 12th December, 2019.