Exploring the Legal Aspect of Inventorship of AI-Based Inventions in India.
Human beings are considered the most intelligent class on the earth. Human beings have always tried to excel in technological aspects so as to make life easier to survive. In the quest of attaining this objective, there is a constant attempt to develop a substitute for the human brain. Such simulation of intelligent human behavior in software is known as artificial intelligence. AI, which was considered part of sci-fi novels has turned out to be a reality with the advancement of technology and the aplomb efforts of human beings.
With the advancement and growth of AI, innovations that would otherwise be impossible through human efforts alone have become possible. To motivate and accelerate the development of AI, patent rights are granted to the inventor in order to ensure the legal security for the protection of AI. Earlier, AI is regarded to contribute only to the invention process without assisting the conception of the invention. But, in recent times scientist envisions that along with assisting in the invention process AI is also capable of creating outputs that would require patentable rights in the future. However, the present patent law is uncertain regarding the validity of computer-generated inventions and ever-advancing AI technology.
In this regard, the fundamental tenets of intellectual property rights must be examined. This article aims to study the concept of AI and the legal aspects of granting patents to AI.
1- AI-generated inventions require patent protection
The basic objective of the patent system is to revitalize research and development investment with reasonable certainty to the innovators. Persistence of uncertainty with regard to the enforceability and validity of a patent can weaken its market value. This will affect the earning capacity of the patent right holder to extract value through licensing. It would not be correct to state that patents un-conditionally protect the holder's investments and guarantee their validity if challenged. But they do pace up innovation and research and disseminate knowledge- the very objective of the patent mechanism is intended to serve.
2- Patentability criteria
The term patent relates to granting exclusive rights to the inventor for a fixed period of time and after the expiry of the term, it passes to the public domain. Patentability criteria differ from country to country with respect to law of the land. Trade-secret aspects on intellectual property (TRIPS) is an international legal document that governs intellectual property rights.
According to TRIPS, to claim patent protection an invention must be new and involve innovative steps, and must be capable of industrial application. In India, the patents act 1970 regulates the legal aspects of a patent and determines whether a particular product or process is eligible for a patent or not. To fulfill the criteria for patentability the particular invention must consist of novelty, inventive step, and industrial application. Section 3(k) of the patents act provides for the patentability of software inventions. It is to be noted that section 3(k) specifically bars the patentability of a computer program per se. however; in general practice, software inventions are patentable if-
i.) The software invention provides practical application and solution to technical problems or an improved technical effect of the underlying software.
ii.) There is technical development in the invention over the existing software applications.
3- Patent law issue with respect to AI
AI is the basic tool of technological and scientific work and creating monopolies on such work through patents would result in hindrance and obstruction of innovation. It has been asserted that a patent should not be granted on works which does not involve any inventive steps and is mere replication of human work. On the same note, another major concern with AI is that it will drastically reduce the labour force participation in the long run. This will increase the risk of wage gaps and economic instability.
4- Issue of Inventorship for AI-generated inventions-
The major conundrum related to the legal aspect of AI is whether the law requires the creator of the AI is named as the inventor or should law permit an AI application to be named as the inventor. The law recognizes that any person who invents is an inventor. However, the notion that AI can invent raises the pertinent question of relating to the ownership concept. One of the tests to determine the ownership is the right to sue or be sued on the particular thing owned. Such right to sue or be sued is provided to corporates under the “legal entity” concept due to the wide-ranging consequences that such corporations can face. Similarly, this precedent can serve to grant legal entity concepts to AI as well. This can be understood from an illustrative example- Company A develops an AI machine, which it sells to company B. Company B operates that AI Machine on resources owned by company C, such as serving in software cloud computing.
Company B also obtains data from company D that is used to upgrade and train the skills of AI machines. After training the AI machine produces inventions. The pertinent question which arises is “who is the inventor”? The present law on patents requires that the formation of ideas must take place in mind. But if all the conception of ideas takes place in the mind of AI, it should be regarded as a person to be listed as an inventor.
Also, there exists a distinction between AI and corporate. AI may be actually independent, but corporate are fictitiously independent and can be held accountable via their stakeholders. One plausible solution in this regard can be granted a specific bundle of rights from those actually provided to legal persons. However, concrete steps in this regard are yet to see.
In the same parlance, considering algorithmic or computational inventions as patentable and providing AI legal status as an inventor would require AI to be treated as a legal person. Being a legal person AI would be subject to rights and liabilities that can arise from the status of being a legal entity. The second alternative in this regard can be, not providing a listing of any inventor. But this can only be possible if the patent legal framework is designed in a manner that patents can be granted to AI without listing the inventor. In adopting this alternative sufficient legal protection have to be drafted which provides incentives and impetus to the creator of AI.
5- Patent infringement liability of AI
i.) Accountability- Patent infringement liability by AI is another pertinent issue that needs to be discussed. A patent grants an exclusive right to the inventor to use or sell the invention. Liability for infringement will arise when someone without permission sells or use such an invention. In case of infringement, the infringer has to pay compensation for damages caused to the third party. But the relevant question in respect of AI will arise as to who be will liable as the infringer.
According to the European Parliament report of 2020, AI cannot be held liable for the loss caused to the third party. Rather, the human agent behind such cause has to be traced such as the manufacturer, operator or the user-provided human agent could have reasonably foreseen the harm that AI will cause. All this is relevant because no one is held liable for patent infringement by AI, it will encourage using AI for infringement of laws.
ii.) Assessment of infringement liability- The next pertinent question arises as to how should liability for patent infringement by an independent AI be handled? One possible aspect, in this case, is the inclusion of an insurance system in which fund is collected for paying damages for infringement. The other aspect would be to hold AI itself responsible. This can be achieved by granting the status of a legal person to AI.
Also, it is relevant to assess the quantum of liability that is to be borne by AI. In the European parliament report of 2020, it was asserted that future legislation should include statutes holding non-humans liable as well. If the liability for infringement is to be attributed to a human agent, then such liability should be in proportion to the extent of authority delegated to AI. But if the AI is to be held accountable, such liability has to be assessed in the same manner as in the case of a corporate entity.
Another option is to have a contractual agreement in case the infringement occurs as it provides an anticipated solution. The aggrieved party would be indemnified and becomes eligible to receive damages according to the relevant clause.
The legislative enactments in respect of different legal aspects of AI will have deep impacts on innovations, the economy, and society. It is pertinent to cautiously evaluate the present yardstick on patent-eligible subject matter to ascertain any material negative impact on AI-driven technologies. The question of whether inventions independently discovered by AI should be protected has to be evaluated in light of its effects-positive or negative emanating from it. In case AI-generated inventions are eligible for patent rights it has to be ascertained whether the inventorship is attributed to AI that generated the inventive ideas. Lastly, patent law with respect to AI must seek to achieve social and economic welfare as a whole.
1. Muskan Saxena, Patenting AI and its Legal Implications, Centre for Intellectual Property Rights Research and Advocacy Nlsiu, (2021). https://iprlawindia.org/wp-content/uploads/2021/04/Muskan-Saxena.pdf.
2. Diamond v. Chakrabarty, 447 U.S. 303 (1980).
3. Frank A. DeCosta, Intellectual Property Protection for Artificial Intelligence,WJIP(2017).https://www.finnegan.com/en/insights/articles/intellectual-property protection-for-artificial-intelligence.html.
4. Pallavi Gupta, Artificial Intelligence: Legal Challenge in India, (2019). file:///C:/Users/prateek-mob/Downloads/artificialintelligence.pdf.
6. Roger C. Schank, what is AI, anyway?, 8(4) AAAIM, Pages 58 – 65, 2001
The article first published on Lexology.com and the same can be accessed here.