Challenges in Patenting AI innovations

  1. Introduction to Artificial Intelligence

Artificial Intelligence or AI is an aspect of computer science which comprises of artificial   neural networks. These are basically computational models which produce output similar to 

Essentially, it is  “Deep supervised machine learning.[1] This fundamentally means that the machine does not require a step by step instruction to function, it educates and trains itself using the patterns in the data it receives. Once it does this, it makes informed decisions just like human beings would.[2] Therefore, it learns cognitively, making it more efficient to process, analyse and handle a lot of data. Similar to humans, the AI’s growth  too centres around capabilities of thinking, learning, memory and perception. With how smart the AI systems are turning, they now assist humans in recognising patterns which were not detected previously. AI as a term is basically used to define tasks which are performed by machine with lack of human intervention.[3]

  1. Issues with patenting

Owing to AI’s component of deep learning and machine learning, it is capable of teaching itself numerous things and activities, thereby making it efficient enough to invent products and processes. Similar to the cognitive abilities of humans, AI’s too are equipped to create inventions, similar to those created by humans. This leads to the interaction patent law has with systems of artificial intelligence, giving rise to several questions and dilemmas. 

Under patent laws, the ‘inventor’ of a process or a product is given exclusive rights over their invention. Such product or process should be such that it is done in a ‘novel’ manner, distinguishing it from the previously existing inventions. Moreover, such an invention must offer a new solution to pre-existing technical issues.[4] The patent holder can further extend the right to use such an invention(for a limited period of time), and also restrict others from creating or selling of such patented process or product. Therefore, the right given to the patent holder are of a nature that it “legitimizes the creation of a monopoly for the benefit of the original inventor.”[5]

There are three criteria under Patent laws which must be fulfilled for an invention to be eligible for patenting. Firstly, the invention must be ‘novel’. Secondly, it must have an ‘inventive step’ and thirdly, it must have some ‘industrial application’.[6]

To satisfy the criteria of novelty, it is imperative of the AI system to assess all the prior art at the invention stage to determine whether what has been created is novel or not, whether it has a creative component attached to it and whether such an invention can be easily anticipated or not. Although the human scientists can input data regarding the prior art into the AI machine however, it raises the question as to whether the AI system is truly independent or not. Such intervention by the human does not make the invention completely the AI’s work. Further, doubts regarding AI’s judgement arises, whether it would capable enough to determine which invention is novel and which is not.[7]

The criteria of the invention having an inventive step would be more difficult for the AI to determine given its limited capabilities at the moment. At present, certain objectives are set and the AI system is simply required to achieve those in a given period of time. Therefore, having human like intelligence to have an inventive step for a process of a product based on pre-existing data and information, which is also unknown to persons who are skilled in that form of art, seems difficult to attain.[8] Moreover, AIs function on mechanical programs and algorithms however, the same are not eligible for patenting. Like in Bilsk v. Kappos,[9] the court denied patent rights to the AI simply because the programs created by it were not inventive but mechanical. Similarly, under the patent laws of India, an explicit exclusion of computer programmes, algorithms, business or mechanical methods etc. is done. Extending such patent protection to these programmes would be redundant as is, given their nature of constantly being updated and revised. Furthermore, expression which have an inventive concept behind them cannot be protected under copyrights laws. Such inventions of programmes or algorithms need to be new for the person who are skilled in that art. However, as technology is progressing and AI’s intelligence is expanding, it is unlikely for AI systems to create inventions unknown to other AI systems, therefore such an invention would be excluded under the threshold of it being a prior public knowledge. Moreover, with how fast AI’s intelligence grasps and processes data, even if the AI is protected, the other systems would likely be at par with its invention or something better in no time making it difficult to patent so many applications.[10]

Now, inventions can be understood by taking the case of Townsend v. Smith.[11] In this case the courts said that a valid invention would be such that goes through the stage of ‘conception’, meaning that the inventor conceives the idea of such an invention in its mind before making good of the same. However, an AI system may not be capable of such conception unlike the human mind.[12] Further, under the Patent laws of the United States, there is a clear exclusion of machine or beings other than humans from their definition of ‘inventor’. According to US laws, an inventor has to be an individual or a set of individuals who creates the invention or discovers the subject matter of the same.[13] Even in the European Patent Office (EUIPO, two patents were rejected in January 2020 for “Food Container”,[14] and “Devices and Methods for Attracting Enhanced Attention”[15]. The office upheld that such patent applications “do not meet the requirement of the European Patent Convention (EPC) that an inventor designated in the application has to be a human being, not a machine.” These application were made by “Artificial Inventor Project”, which had filed for such AI created invention applications in 7 other nations, even the UK. The same was rejected by the UK Patents Office on the grounds that “there appears to be no law that allows for the transfer of ownership of the invention from the inventor to the owner in this case, as the inventor itself cannot hold property.”[16] Although, AI systems cannot be given such patent rights which are intangible in nature, neither can they be held accountable for it however, even the human owner of such system cannot be listed as the inventor as it is a punishable offence in most countries to be a patent holder without making contributions to the same. In the re Hardee case, the court held that “The threshold question in determining inventorship is who conceived the invention.4 One must contribute to the conception to be an inventor.”[17]

Since the human is the one which creates the AI and will probably assist it in distinguishing its invention from prior art by instilling the necessary data, the argument of collaborative invention comes into play. However, an AI application cannot be termed as a ‘legal personality’ in most jurisdictions. The only other way is to define AI system as an ‘electronic person’, thereby distinguishing it from the ‘natural person’ under revised patent laws to grant such rights.[18]

Furthermore, patent protection is provided as an incentive to create innovation among humans. However, this cannot be a motivation for AI systems. Human inventors have an attachment to their invention which the AI lacks, making it incapable of exercising the exclusive rights provided under patent protection in an effective manner. This would completely defeat the purpose of patent rights as a whole.[19]


[1] Allen, Gregory. “Department of Defense Joint AI Center-Understanding AI Technology.” The Official Site of the Department of Defense Joint Artificial Intelligence Center, Apr. 2020. 

[2] Jelinek, F. “Statistical Methods for Speech Recognition.” 1998. 

[3] Kokane, Sonali. “The Intellectual Property Rights of Artificial Intelligence-Based Inventions.” Journal of Scientific Research, vol. 65, no. 02, 2021, pp. 116–119., https://doi.org/10.37398/jsr.2021.650223. 

[4]  “What Is a Patent?” Patents, n.d., http://www.wipo.int/patents/en/. 

[5] Tripathi, Swapnil, and Chandni Ghatak. “Artificial Intelligence and Intellectual Property Law.” Christ University Law Journal, 2018. 

[6] The Patents Act, 1970.

[7] Yu, Ronald. “Should an Artificial Intelligence Be Allowed to Get a Patent?” Robohub, n.d., http://robohub.org/should-an-artificial-intelligence-be-allowed-to-get-a-patent/.

[8] Yu, Ronald. “Should an Artificial Intelligence Be Allowed to Get a Patent?” Robohub, n.d., http://robohub.org/should-an-artificial-intelligence-be-allowed-to-get-a-patent/.

[9] Bilsk v Kappos, 561 U.S. 593, 2010.

[10] Lazaro, Lynn. “Artificial Intelligence in the World of IP – Intellectual Property – India.” Welcome to Mondaq, Kochhar & Co., 10 Feb. 2020, https://www.mondaq.com/india/patent/892134/artificial-intelligence-in-the-world-of-ip.

[11] Townsend v Smith, 36 F.2d 292,293, 1929.

[12] Tripathi, Swapnil, and Chandni Ghatak. “Artificial Intelligence and Intellectual Property Law.” Christ University Law Journal, 2018. 

[13] Lohr, Jason. “Http://Www.limegreenipnews.com/2016/07/Artificial-Intelligence-Drives-New-Thinking-on-Patent-Rights.” LIME GREEN, 2016, http://www.limegreenipnews.com/2016/07/artificial-intelligence-drives-new-thinking-on-patent-rights/. . 

[14] Patent Application # EP3564144.

[15] Patent Application # EP3563896.

[16] Lazaro, Lynn. “Artificial Intelligence in the World of IP – Intellectual Property – India.” Welcome to Mondaq, Kochhar & Co., 10 Feb. 2020, https://www.mondaq.com/india/patent/892134/artificial-intelligence-in-the-world-of-ip.  

[17] Re Hardee, 223 USPQ 1122, 1123, 1984.

[18] Lazaro, Lynn. “Artificial Intelligence in the World of IP – Intellectual Property – India.” Welcome to Mondaq, Kochhar & Co., 10 Feb. 2020, https://www.mondaq.com/india/patent/892134/artificial-intelligence-in-the-world-of-ip.

[19] Tripathi, Swapnil, and Chandni Ghatak. “Artificial Intelligence and Intellectual Property Law.” Christ University Law Journal, 2018. 

Aishwarya Says:

I have always been against Glorifying Over Work and therefore, in the year 2021, I have decided to launch this campaign “Balancing Life”and talk about this wrong practice, that we have been following since last few years. I will be talking to and interviewing around 1 lakh people in the coming 2021 and publish their interview regarding their opinion on glamourising Over Work.

If you are interested in participating in the same, do let me know.

Do follow me on FacebookTwitter  Youtube and Instagram.

The copyright of this Article belongs exclusively to Ms. Aishwarya Sandeep. Reproduction of the same, without permission will amount to Copyright Infringement. Appropriate Legal Action under the Indian Laws will be taken.

If you would also like to contribute to my website, then do share your articles or poems at secondinnings.hr@gmail.com

In the year 2021, we wrote about 1000 Inspirational Women In India, in the year 2022, we would be featuring 5000 Start Up Stories.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.