From ‘who’ to ‘what’: Proposed Paradigm shift in Patent Law

By Aditya Gupta


“The short-term impact of AI depends on who controls it, the long-term impact depends on whether it can be controlled at all” – Stephen Hawking[1]


As abstract and far-fetched as it may sound, two patents have been filed before the US Patent and Trademark Office, the European Patent Office, and the UK Intellectual Property Office, the credit of which has been given to a machine.[2] The team responsible for the patent applications have also registered and launched a website for detailing the viability and requirement of recognizing machines as inventors.[3] The website in question,, contains an excellent collection of arguments used by the pedagogy in support of accepting Artificial Intelligence [“AI”] as an inventor.

Owing to the ubiquity and importance of the field of innovation, academia has been ablaze with questions related to patentability of the AI inventions[4], the shortcomings of the Trade Secret protection regime for protecting AI-based inventions[5], and the development of Copyright Law to accommodate AI created copyrightable subject matter.[6] While some consider the notion of AI as an inventor of patents as ‘stupidity’[7], the filing of the present patent applications are perplexing issues for the legislature and a new scope of assessment and criticism for the academia. 

The Proponents: A Case For an Artificial Inventor as an Inventor

At the forefront of the issue is Prof. Ryan Abbott, who argues that patent regimes should be globally overhauled to allow ‘sentient machines’ to be recognised as inventors. While the majority of the arguments espoused by DABUS proponents are based on various policy considerations, Prof. Abbott has borrowed a legal interpretation from one of his publications;[8] the argument rests on the premise that American Patent Law, while defining the import of the term ‘Inventor’, uses the word ‘Individual’ [9].  Since an Inventor has been defined as an Individual, DABUS can be referred to as an Inventor owing to the fact that the present inventions have been independently created by DABUS. Prof. Abbott further posits that since there is no explicit bar against an AI being recognised as an individual, the courts should assume a liberal interpretation of the law and allow a Computer Programme to be circumscribed within the definition of the term inventor.[10]

The Critics: The Reservations for AI as an Inventor

The critics or the exponents of the idea have mostly relied on two arguments. Firstly, they do not comply with the definition of individual as espoused by Prof. Abbott; for instance, Dr Shlomit Ravid, professor of law at the Yale Law School, relies on § 101 of the America Invents Act and underlines the usage of the term “whoever”, while asking the question of whether an AI system can be designated as a ‘who.’ She further adds that the American legislative framework has been drafted in a manner which requires human interaction for being designated as an inventor for a patent.[11]

Secondly, the case where a monkey (Naruto) was not allowed to own copyright over the photos he took of himself. Analogically, the authors argue that when a sentient being such as a monkey was not allowed to own the copyright in his creation, the chances of an AI being designated as an inventor are very bleak.[12]

While the latter argument does not hold much merit, the former assumes significance. Prof. Abbott’s interpretation of the term ‘individual,’ is the primary argument based on interpretation of the law. If the academia can decimate the applicants’ argument, they probably make the job of the United States Patent and Trademark Office and the courts much easier. The fact remains that mere recognition of AI as an ‘individual’ would not solve any problem. In fact, it would open a Pandora’s box of compliances, amendments and litigations within the patent, legal and managerial framework. While the definition of Inventor might be favourable to the present case, the scheme of the patent law within the American jurisdiction provides for various other provisions which would not be as favourable. For instance, §102 of the patent law, while elucidating the qualification required for patent protection, uses the term ‘person,’ which is arguably different in its meaning to the term ‘individual.’

The Situation in India: Proponent or Exponent?

The determination of the stance in India on the subject rests on S. 6 of the Patents Act and the interpretation of the terms involved therein. S. 6 prescribes that only the ‘true and first inventor’ and other persons deriving right therefrom are entitled to apply for the grant of patents. The term ‘true and first inventor’ has been accorded an exclusionary definition[13] and therefore it can be safely submitted that an AI programme can fit within the contours of a ‘true and first inventor’. Although, it has been qualified by the term ‘person’, which has been defined to include the government. Judicial precedent on the subject dictates that a company or a firm cannot be named as a true and first inventor in a patent application.[14] Drawing a corollary here, S. 6 requires a particular threshold of human control and input in the invention process. In fact, N.R. Subbaram in his book ‘Patent Law: Practices and Procedures’ has gone ahead to state that “the true and first inventor(s) should always be a natural person, namely, a human being, because only human beings are vested with the power of intelligence and capable of providing intellectual inputs.”[15] Although published in 2007 it can be argued that this instruction does not hold relevance in the contemporary technology-driven market.


The interpretation of the term ‘person,’ in light of the legislative instruction forms the cornerstone of the discussion within the Indian framework. If an artificial intelligence can assume the status of a person, it might as well be appreciated as the inventor of a patent. Although, the fact remains that a country which does not allow companies to be referred to as inventors is highly unlikely to admit the argument that a machine should be accorded the status. The present situation dictates that the debate around granting inventorship to AI programs circles around the various policy concerns which have been extracted by AI advocates. Although, the fact remains that patent law in its present structure and form is not capable of such wide interpretation. The efficacy of such an overhaul has to be established on the pedestal of viability. The proponents and exponents have both cited policy concerns and unless and until the proponents can posit some reasons for such an expansive legislative overhaul, the status quo should be maintained.

As for India, the fact remains that S. 2(1)(y) of the Patents Act was drafted in 1970 and it is highly improbable that the drafters of the statute would have foreseen a situation where a computer programme could have been identified as a ‘True and First Inventor.’ If the legislative intent behind the definition of the term ‘person’ is to be assumed to give due credit to the individual who has invented the subject invention, it is highly unlikely that this underlying intent shall be extended to include AI programs, singularly owing to the fact that the statute was drafted at a time where the present application was virtually ‘science fiction’.  

The author is a 9th Semester student currently studying at the National University of Study and Research in Law (NUSRL), Ranchi.

[1] Stephen Hawking, Stuart Russell et. al., Stephen Hawking: ‘Transcendence looks at the implications of artificial intelligence – but are we taking AI seriously enough?’ Independent (May 1, 2014),

[2] Martin Coulter, Patent agencies challenged to accept AI inventor, Financial Times (Aug. 1, 2019),

[3] Leo Kelion, AI system ‘should be recognised as inventor’, BBC (Aug. 1, 2019),

[4] Clark D. Asay, Artificial Stupidity, 61Willian and Mary L. Rev., (2020) Forthcoming.

[5] Jeanne Fromer, Machines as the New Oompa-Loompas: Trade Secrecy, the Cloud, Machine Learning, and Automation, 94 NYU L. Rev., (2019) Forthcoming.  

[6] Andrez Guadamuz, Artificial Intelligence and Copyright, WIPO Magazine, October, 2017 No. 5 at 14.

[7] Alex Castro, This Is When AI’s Top Researchers Think Artificial General Intelligence Will Be Achieved, The Verge (Nov. 27, 2018), as cited in supra note 4 at 5.

[8] Ryan Abbott, I Think, Therefore I Invent Creative Computers and the Future of Patent Law, Bos. Coll. L. Rev. 1079, 1096-1100 (2016); Also see: Ryan Abbott, Patenting the Output of Autonomously Inventive Machines, 10 Landslide, (2017);

[9] 35 U.S.C. § 100(f): defines the term inventor is as: “the individual or, if a joint invention, the individuals collectively who invented or discovered the subject matter of the invention.

[10] Benjamin Ford, Artificial Intelligence Inventor Asks If ‘WHO’ Can Be an Inventor Is the Wrong Question? IPWatchdog (August 5, 2019),

[11] Shlomit Yanisky Ravid & Xiaoqiong Liu, When Artificial Intelligence Systems Produce Inventions: An Alternative Model for Patent Law at the 3A Era, 39 Cardozo Law Review 2231 (2018).

[12] Id.

[13] The Patents Act, 1970, No. 39, Acts of Parliament, 1970, S. 2(1)(y).

[14] 23 Halsbury Laws of India, Intellectual Property Rights 69 (Lexis Nexis,1st Ed., 2009); Also see: V.B. Mohammed Ibrahim v. Alfred Schafranek, AIR 1960 Mys 173 (DB), Darius Rutton Kavasmaneck v. Gharda Chemicals Ltd. and Ors.2015 SCC Online Bom 4813.

[15] Subbaram’s Patent Law; Practices and Procedures 72 (Lexis Nexis 2nd Ed., 2007).