Implications Of Artificial Intelligence On Patent Law

Author: C. Srinidhi
Junior Research Fellow
Osmania University

Artificial Intelligence (AI) refers to computer-related inventions which invent machines that work and think like humans. It creates machines which emulate human intelligence. Artificial Intelligence which was once dismissed as merely being science fiction has now become reality having an impact on all industries ranging from healthcare to education. AI is estimated to contribute 30 trillion to the global output by the year 2030. Investment in R&D of AI is being done not only by private organizations but also by several governments such as China, the United States and the European Union. This increased boost in investment has led to an increase in intellectual property involving AI. Patent applications related to AI have been seen on the rise across all major patent offices of the world. It has been observed that the patent offices are unsure of how to deal with these applications related to AI because patent law as it stands today is not equipped to regulate AI. Several important questions have arisen concerning subject matter criteria, inventorship and ownership status to non-humans, infringement liability. As AI transitions from ‘narrow AI’ where the ability of AI is limited to performing a specific task to ‘strong AI’ where AI is able to generate inventions autonomously independent of humans these questions become pertinent. For instance, if an invention is generated by AI independently, should the machine be designated as the inventor of the human who made that machine. With regard to AI-assisted inventions, should AI be designated as a joint inventor, if yes, how to determine the contribution of AI and the human inventor.

These rapid technological advances in AI are expected to disrupt the current patent system. AI will challenge the core legal standards of patent law; subject matter eligibility, novelty, inventive step and utility.[1] In order to keep up with AI, the patent law must be updated accordingly. Even though AI is said to have enormous benefits to society as a whole, unregulated AI will have serious social, legal and ethical complications. Elon Musk, CEO of Tesla and SpaceX cautions “AI is a rare case where I think we need to be proactive in regulation than be reactive”[2] He also suggests that there should be a regulating mechanism to oversee the development of AI.[3] This nature of AI demands inquiry into what challenges does AI pose to the current patent system.

When it comes to protecting AI with patent law, certain issues arise; whether inventions generated by AI should be accorded patent protection (whether AI needs the incentive to innovate like humans). If yes, to what extent and most importantly, whether AI should be given inventorship status.[4] At its core, patent law is a bargain between society and the inventor. The inventor gets a monopoly on his invention for a specific period of time in exchange for full disclosure of his invention. This monopoly granted to the inventor acts as an incentive to innovate. In order to get patent protection, an invention must satisfy the criteria of subject matter eligibility, novelty, inventive step and utility. AI collides with each of the requirement of patent law. Patent law excludes ‘mental processes’ and ‘abstract ideas’ from being patented, however, AI exactly automates ‘mental processes’ which can be done by the human mind.[5] Currently, AI is capable of completing the inventing and patenting process on its own without human intervention. With some AIs it is difficult to fully disclose the invention. This is because; when an AI is being created it is ‘trained’ by the inventor by providing feedback on the output it delivers by processing a set of ‘training data’. The software produced by this training is a result of “jumble of weights and interconnections between nodes in a neural network or some similarly human-illegible chunk of math.”[6] This means that the AI does not provide the inventor with any logic or reason behind its decisions. Hence, the inventor himself does not know how and why the AI has created something. This will make it difficult to fully disclose the workings of the invention. Another question which arises is related to ‘inventorship’; whether the AI can be given the status of an inventor or does patent law require a ‘human agent’. Determining ‘inventive step’ or ‘non-obviousness’ is also problematic in AI because ‘what is obvious into the super-intelligent AI’? Also, the question of what should be the criteria for determining ‘person skilled in the art’ also arises.

The current patent system is not ready for the onslaught of AI. If AI is attempted to be protected under patent law, it will surely disrupt the patent system. It is extremely important to proactively legislate a sui generis system or amend the current patent system to protect AI.

[1] White Paper on Artificial Intelligence Collides with Patent Law, World Economic Forum, April 2018

[2] https://www.cnbc.com/2017/07/17/elon-musk-robots-will-be-able-to-do-everything-better-than-us.html.

He also said “If you ask ‘What harm can a deep intelligence in the network do?’ Well, it can start a war by doing fake news and spoofing email accounts and doing fake press releases and by manipulating information”

[3] Id.

[4] Supra at 1.

[5] Alice Corp. v. CLS Bank Int’l, 134 S. Ct. 2347, 2355 (2014) “courts should generally exclude patent claims directed to mental process”, “in the human mind” or by “a human using a pen and paper” with the limited exception for claims that specifically provide ways to achieve technological improvements over the tasks previously performed by people (e.g. containing an “inventive concept”)

subject matter that could be performed through an “ordinary

[6] Stacy Rush, Artificial intelligence’s “black box” decision-making presents challenges for AI and machine learning innovators who want to file for patents, Canadian Lawyer Magazine, 27^th November 2017. https://www.canadianlawyermag.com/news/opinion/the-challenges-of-patenting-artificial-intelligence/274698. Last Accessed on 29th September 2019.