Abstract
The rise of Artificial Intelligence (AI) brings both promise and skepticism. Skepticism arises from the unpredictable consequences and the complexities of AI technologies, leading to concerns about misuse and potential negative impacts on society. Despite this, encouraging AI development offers significant benefits, from data-driven decision-making to accelerated scientific breakthroughs. However, patentability challenges emerge, especially in India, where AI-related inventions face scrutiny under existing laws.
The intersection of AI and patent law, particularly in India, poses challenges, with Section 3(k) of The Patents Act, 1970 that excludes certain AI-related inventions from patentability. Recent legal cases, such as Microsoft Technology Licensing, LLC v. Assistant Controller of Patents and Designs and Ferid Allani vs Union of India highlight the evolving interpretation of “computer programme per se”. The Delhi High Court emphasizes the need for a technical contribution beyond the program itself for patent eligibility.
Defining a patent eligibility framework for AI-related inventions becomes crucial to strike a balance between inventor rights and the public interest. The challenge lies in adapting to the dynamic nature of technology, clarifying concepts like “technical contribution” and “technical effect” to ensure continued innovation and societal welfare. AI’s rapid growth and integration into research raise challenges without clear rules for protecting inventions and determining ownership.
In James Cameron’s film “The Terminator”, featuring Arnold Schwarzenegger, audiences witnessed the malevolent Artificial Intelligence (AI) from the year 2029. This diabolical AI sends a meticulously designed Terminator back to 1984 with the explicit mission of killing a woman who will give birth to a future threat to the very existence of robots.
The deployment of AI systems may have unintended consequences that are difficult to predict. The complexity of AI technologies contributes to skepticism among people who may not fully comprehend how these systems work. A lack of public understanding can lead to mistrust and apprehension. The power and capabilities of advanced AI raise concerns about the potential for misuse, whether intentionally or unintentionally. This includes scenarios where AI is used for malicious purposes or to manipulate information. This uncertainty often contributes to skepticism, as people worry about unforeseen negative impacts on society.
However, encouraging the development and use of AI can bring about several potential benefits and advancements. AI systems can analyze vast amounts of data quickly and make data-driven decisions, potentially leading to more informed and accurate choices in various fields such as healthcare, finance, and business. AI can accelerate scientific research by analyzing vast datasets, simulating experiments, and identifying patterns that may not be apparent to human researchers. This can expedite breakthroughs in various scientific fields. Encouraging the responsible development of AI promotes a future where humans and AI systems work collaboratively, leveraging the strengths of both for mutual benefit.
Traditionally, it has been believed that humans, with their intelligence, can create novel and innovative work that can be acknowledged and registered as intellectual property. This belief is also evident in the laws made by legislators for intellectual property.
On the other hand, imagine a machine or software capable of understanding information fed into it, generating intelligence that is vastly superior, precise, and accessible to all. Shouldn’t such intelligence be granted similar rights as intellectual property created by humans? Shouldn’t the creator of such a machine be recognized for the work it produces, or should the machine itself not be acknowledged for autonomously developing work without human intervention?
The advent of AI innovations has disrupted the conventional standard of patentability. This emerging technology presents novel challenges to the patent system, necessitating a concise examination of specific issues pertinent from a patent perspective.
There has been an increased filing of patent applications on inventions generated by AI, not only in India but at several patent offices around the world. Fundamentally, most AI generated inventions relate to software codes that direct the AI system in executing actions, making decisions, and producing outputs according to predefined commands. When numerous such commands are assembled into a computer program and the AI system is given data as input, the software can generate outputs without additional instructions.
Hence, the more significant question revolves around how to scrutinize patent applications for AI-related inventions, including both AI inventions and AI-created inventions.
AI Inventions in India
In India, the fundamental requirements for an invention are laid down in Section 2(1)(j) of the Patents Act, 1970.
Section 2(1) (j) “invention” means a new product or process involving an inventive step and capable of industrial application;[1]
Like any other invention, to be patentable, AI generated inventions must also satisfy the fundamental requirements of novelty, inventive step, and industrial application as under section 2(1)(j) of the Patents Act, 1970.
The Act also furnishes a non-exhaustive list of “what are not inventions” under Section 3 of the Patents Act, 1970, outlining subject matter and activities that are excluded from patentability.
More specifically, Sections 3(k), 3(m), and 3(n) of the Indian Patents Act, 1970 relates to inventions concerning AI.
Section 3 (k) a mathematical or business method or a computer programme per se or algorithms;
Section 3 (m) a mere scheme or rule or method of performing mental act or method of playing game;
Section 3 (n) a presentation of information;[2]
A recent case of Microsoft Technology Licensing, LLC v. Assistant Controller of Patents and Designs [2023: DHC: 3342][3] involved interpretation of words ‘per se’ as in Section 3(k) of the Patents Act, 1970 which lists ‘computer programs per se’ as not patentable.
In this case, on 07/11/2003 Microsoft filed a patent application (1373/DEL/2003) for the invention titled, “Methods and Systems for Authentication of a User for Sub-locations of a Network Location.” The Patent Office raised objections in its First Examination Report (FER) on the grounds of non-patentability under Section 3(k) of the Act (computer programme per se) among others. This was followed by the hearing where Microsoft submitted its arguments, but the Controller remained convinced and issued the order rejecting Microsoft’s patent application. Microsoft appealed to the High Court of Delhi, contesting the Controller’s decision.
The Court noted that the term “computer programme” was inserted into Section 3 as a non patentable invention only in 1999, through the Patents (Second Amendment) Bill, 1999. In 2001, the Joint Parliamentary Committee report recommended the insertion of the word “per se” in conjunction with “computer program” in Section 3(k), explaining that the intention here is not to reject computer programs for grant of a patent if they are inventions, but to refuse the patent to the computer programmes as such. The Patents (Second Amendment) Bill, 1999, later enacted as The Patents (Amendment) Act, 2002, having the Section 3(k) with the phrase “computer programme per se.”
The Court noted that the intent of the amendment was to allow the grant of patents to computer related inventions that involve a novel hardware component or provide a technical contribution to the prior art(s) beyond the programme itself.
The High Court, in its decision dated 15th May, 2023 was of the view that if the subject matter is implemented on a general-purpose computer, but results in a technical effect that improves the computer system’s functionality and effectiveness, the claimed invention cannot be rejected on ground of non-patentability as ‘computer programme per se’.
In another case, Ferid Allani vs Union of India[4] [WP(C) 7 of 2014], the Court ruled that adding the term ‘per se’ to Section 3(k) aimed to clarify that authentic inventions based on computer programs should not be denied grant of patents. The Court highlighted the need to delve into the aspect of ‘technical effect’ and ‘technical advancement’, for determining the patentability of computer related inventions.
In other words, ‘method claims’ within computer programme patents may qualify for patentability if they include a technological advancement, offer a technical solution to a technical problem, and exhibit an enhanced technical effect on the underlying software.
Therefore, understanding the technical contribution of a mathematical method employed by AI involves assessing whether the method, within the context of the invention, serves a distinct technical objective. Claims in an AI patent must be functionally confined to this particular technical purpose, whether explicitly stated or implied. Achieving this involves establishing a substantial connection between the technical purpose and the steps of the mathematical method.
AI Inventorship in India
The Patents Act, 1970 explicitly lays down who will be considered as the inventor for AI based inventions that require human intervention. Section 6 of the Patents Act clearly outlines the persons entitled to apply for patents. This section enables persons to file for a patent for an invention, subject to specified provisions. These persons include the actual and first inventor, the inventor’s assignee, and the legal representative of a deceased person who had been entitled to file the application prior to their death. The application can be made alone or jointly as per this section.5]
Section 134 excludes nationals from countries not providing for reciprocity from applying for grant of patents.
Thus, an application for a patent for any invention can be made only by the true and first ‘inventor’ of the invention or his assignee and that the ‘person’ is the true and first inventor. The act is silent on the inventorship of AI, in case of AI-generated inventions.
Section 2(1)(p) of the Patents Act defines ‘patentee’ as the ‘person’ entered on the patent office register as the grantee or owner of the patent. This seems to imply that ‘person’, here, refers to an individual, a real person (a natural person). On the other hand, the argument comes up while looking into Section 2(1)(s) that defines ‘person’ to include the government, which is a non-natural entity. However, neither Section 2(s) nor Section 2(t) explicitly clarifies if AI can be considered as a person.
Section 2 (p) “patentee” means the person for the time being entered on the register as the grantee or proprietor of the patent;
Section 2 (s) “person” includes the Government;
Section 2 (t) “person interested” includes a person engaged in, or in promoting, research in the same field as that to which the invention relates;[6]
In a recent application (202017019068) titled, “Food Container and Devices and Methods for Attracting Enhanced Attention”, inventor being an AI, Device for Autonomous Bootstrapping of Unified Sentience (DABUS), the Controller has raised objections against acknowledging AI as an inventor citing the provisions laid in Section 2 and Section 6 of the Patents Act, 1970. Even though the application is still undergoing prosecution, AI as patent holders within the existing legislative framework in India appears infeasible.
Moreover, contemplating AI as an inventor may raise liability concerns. Section 48 of the Patents Act grants the patentee the exclusive right to prevent unauthorized acts by third parties. The crucial question here is whether AI possesses the capability to give consent, and if so, how would one obtain such consent?
Section 48 of the Indian Patent Act grants patentees exclusive rights to prevent third parties from making, using, offering for sale, selling, or importing products in India without their consent. Where the subject matter of the patent is a process, the patentee has the exclusive right to prevent third parties from using the process and, offering for sale, selling, or importing products obtained directly by that process in India under this section.[7]
Addressing liability for an infringing AI poses questions: How would the courts hold an AI accountable for infringement? Would legal responsibility for the AI’s unlawful actions rest with the AI itself, its owner, or its user/operator? If the origin of the unlawful act cannot be linked to a specific human, who bears the liability? These and related concerns are currently under scrutiny, sparking debates on the uncertainties surrounding AI. This extends beyond intellectual property debates, delving into discussions about criminal or civil tort liability.
Perhaps, the legal system could outline specific criteria for attributing liability, considering factors such as the level of autonomy and decision-making capacity of the AI. There can be explicit guidelines on the responsibilities of AI entities, their owners, and users/operators.
Conclusion
Granting patent protection to AI-generated creations is crucial for stimulating innovation. While AI itself may not be motivated by the prospect of obtaining a patent, the availability of such protection will serve as an incentive for individuals involved in the development, ownership, and utilization of AI. Consequently, permitting patents on AI-generated works will foster the advancement of creative AI, leading to increased innovation within society.
Moreover, patents play a pivotal role in encouraging the disclosure of information and the commercialization of socially valuable products. This holds true for AI-generated works, as patents in this domain can effectively achieve these objectives, parallel to their impact on other patented innovations.
Continuing the advancement of AI-related inventions necessitates establishing a patent eligibility framework, offering clear definitions of “technical contribution” and “technical effect”. The ever-evolving technology landscape presents challenges, as these definitions may become outdated in the future. AI is experiencing rapid growth, and it wouldn’t be inaccurate to suggest that it is gradually assuming roles traditionally held by humans. Inventive AI is increasingly integral to research and development. Without clear rules or guidelines, protecting AI-generated inventions poses significant challenges. Hence, there’s an urgent necessity to clarify these concepts, achieving a balance between safeguarding inventors’ rights and fostering public interest and social welfare.
By Bhaskar Dutta, 1st Year LL.B, Faculty of Law, The Maharaja Sayajirao University of Baroda, Vadodara
