Patenting Artificial Intelligence: Legal Implications

Written jointly with Janaki Mandala


The patent system in today’s world is challenged like never before. The criteria of novelty and non-obviousness are becoming increasingly complex to determine with great level of certainty. Today, the software created by Artificial Intelligence (AI) technologies are devising products and services that have not been seen before. The U.S. alone accounts for more than 22,000 AI patent applications.

John McCarthy, a computer scientist, coined the term ‘AI.’ He defined AI as “the science and engineering of making intelligent machines”. AI is a new digital innovation that will have a profound impact on the world. It will create immense technological, economic, and social repercussions, and will transform not only the way we produce and distribute goods and services, but also the way we work and live.

In AI, machines are creators. AI is a subject-matter of myriad patents founded on inventions in machine learning (ML), deep machining, natural language processing (NLP), natural language search, and data analytics. AI is automating existing resources and data to create new inventions for corporates.

There have been over 154,000 AI patents filed globally since 2010. The revenue from AI is set to catapult from $8 billion in 2016 to over $47 billion in 2020. In 2018, IBM inventors were granted about 1,600 AI patents. Microsoft leads the AI patent race, with 697 world-class patents, as of November 2018, considered to have a significant competitive impact. As more organizations embrace AI inventions, the journey of patenting AI inventions keeps gaining rapid momentum.

The moot question is whether the current patent system is able to provide adequate protection for inventions relying on AI without human input and the subsequent legal implication arising out of such non-protection.

Patenting Ai – Issues & Challenges

The development of AI has a very conspicuous offshoot: the protection and controlled use of the invention, or the techniques that encourage such an invention. Legal issues in patent law may include questions as to who would own a patent if the inventor of an AI system discovers a new innovation? Who would be liable if an AI system makes a decision that causes financial damage or injury?

But, can AI be truly protected via the current patent laws? If so, who bags the patent rights – the owner, inventor, programmer, or the AI itself? In case of the eligibility of patent grant to a computer-generated invention, there are three primary probable scenarios:

The patent would be granted to:

  1. The computer
  2. The owner of the computer
  3. Such person(s) who have compiled the hardware in such a way so as to create an invention.

One may also state the AI isn’t eligible to apply for a patent since it isn't a human. The laws, however, recognize the existence of a ‘legal person,’ in addition to a ‘natural person.’ The procedure to obtain a patent for the AI would be possible. But the moot question is: is it really possible to patent a technology that is subject to the society’s dynamic nature? AI is authorized by the inevitability to change and evolve with the changing times and attitudes of humans. It would be a difficult task to patent every single software that promotes a change, even if we forecast the changes that humans, as a society or as intellectual beings, may experience.

So, ultimately, who gets the patent? An inventor has invested his heart and soul in the invention. But, the programmer of the invention has an equally critical role to play. Also, can a person, who buys or lawfully owns the patent, and thereby its technology too, be excluded from granting of the patent? Last but not the least, can the patent be granted to the AI itself? Since the AI is embedded with a programming and function, akin to that of humans, can we consider granting the patent to the AI itself? After all, it now possesses the understanding to distinguish between an individual and a non-individual. Would it be injustice, if the AI was considered an object and not a human clone?

While the argument on the recognition of inventions pertaining to AI is not yet conclusive, the topic has continuously emanated other significant issues. For instance, even if AI were to obtain IP recognition, who would commercialize the exclusive rights? Also, if ownership is accorded to the AI developer as a benefit for effort and investment, why should the developer, who may be involved only during the initial production stage, be rewarded for the final production stage as well? Moreover, if the works produced by AI were to come into public domain, why would developers exert their mental and financial endeavors to develop AI with full intensity?

The outcomes produced through AI are either a consequence of its inherent intelligence or an algorithm. If the functions of the machine are absolutely mechanical, rather than inventive, AI might be regarded as lacking creativity. However, initially, a differentiation between deep-learning — the process in which AI can identify and understand information and data, administered or otherwise — and general algorithms should be made.

The existing laws of any country would not recognize AI as an author or creator of IP. Therefore, AI would not be granted ownership unless it is able to attain a legal status similar to that of inventions by humans. The IP laws of most nations necessitate a rights holder to have legal personhood, something that AI falls short of. In the future, AI might be able to outdo human intelligence and lead mankind to new discoveries, which the laws must be able to protect. Ultimately, if AI is able to establish independent creativity, it could be regarded a potential author, apart from the human author under copyright law. Machines that have the capacity to develop and enhance their ability through incessant learning and training — as against those that operate through a series of algorithmic steps — could be eligible for patent rights.

To claim IP protection for software or inventions based on AI, the following is recommended:

  • Explain hardware (e.g., computer system, server, sensors, etc.) together with AI algorithms in the patent
  • Claim working method/process of device that uses AI
  • Abstain from focusing directly on AI-based programming codes/algorithms.

An issue worth discussing would be that if an AI commits a breach – of law or rules prescribed for the human society – would it be held individually liable for its actions, and prosecuted and penalized? If not, whom would the vicarious liability fall upon?

The World Intellectual Property Organization's definition of IP mentions about ‘creations of the mind’, but does not specify whether it must be the mind of a human or otherwise. But, despite the comprehensive nature of the definition, and myriad laws that support and encourage this change, it remains a Herculean task to identify, let alone grant, the true owner of the rights and benefits of a patent.

The legal implications of AI in India are unknown at present, and the patent regime is silent on the matter. Thus, three main issues stand out from a patent perspective:

  1. Subject-Matter Eligibility: Is AI, as an invention, eligible subject matter?
  2. Creatorship: Who is the true and first creator?
  3. Liability: Who owns the liability for the acts of the AI technology.


In the above scheme of things, it is safe to say the concept of AI, as it keeps evolving, is yet to be in sync with the legal and other aspects of the society. It is yet to gain compatibility with the patent laws on an international level. Apart from encouraging creativity, there should be a concerted effort to regulate and organize the growth of this immense field. The idea of creating a humanoid or the concept of substitution of human intelligence with machines and bots may sound futuristic and brazen, but will inevitably have a profound impact on the mankind as a whole.

The existing patent system is suited to adjudicating the patentability questions related to only the current generation of AI technology. However, the time has come for the Indian government to begin carefully considering how CRIs pertaining to the next generation of AI should be regarded in the patent ecosystem. The law has to be reformed to include AI inventors to expedite innovations. The world must equally take cognizance of the flop side of AI inventions, which may lead to degeneration of human intelligence. It should create a balanced field for the human and AI inventors.

In the future, computers will be responsible for almost all inventions. The creative machines will be smart professionals in their spheres, with unlimited knowledge to create innovations. Unambiguous agreements on IP rights should be established between third-parties to manage risk. Currently, without any specific legislation governing the recognition of AI under IP law, existing challenges can be resolved through a clear agreement between relevant parties, i.e., the AI developer and user, to utilize and commercialize IP created by AI.

Image (c)


More by :  P. Mohan Chandran

Top | Business

Views: 3721      Comments: 0

Name *

Email ID

Comment *
Verification Code*

Can't read? Reload

Please fill the above code for verification.