This article has been written by Ram Kumar, from Chanakya National Law University, Patna.

Abstract

In the last decade, Artificial Intelligence has caused a widespread discussion among academia, business, lawmakers and policymakers for the need to recognize its copyrightability and patentability. This blog post deals with some of the current examples of AI like robot reporters such as Heligrof, Cyborg, Bertie, Next Rembrandt (AI-3D painter), e-David (a painting machine) etc. This paper further deals with the current copyright and patent laws in countries such as US, UK, France, Netherland, Canada, Germany as well as European Commission Directives and Berne Convention. By analyzing these legislations, the author argued how some legislation, like the UK, are more favorable towards recognizing AI-generated works than like US laws. The author finally concludes by recognizing the need to overhaul the existing legal framework by carrying out more economic-based research and at the same time cautious and define the liabilities of AI in case of IPR infringement.

Introduction

Alan Turning first defined Artificial Intelligence (AI) as a machine with a form of intelligence which later popularized by John McCarthy and Marvin Lee Minsky, organizers of Dartmouth Conference in 1956.  AI must be viewed not to replace humans but to augment human beings because humans have too much data to read, understand and analyze due to the explosion of information in the present era. AI has caused a widespread discussion among the policymakers and in public. In the last few years or decade, investment in the field of Artificial Intelligence is skyrocketing. It garnered much attention not only from the public but also from policymakers, judiciary, academia, institutions and business organizations. AI continues to challenge the existing legal framework of Patent and Copyright Laws around the world. AI raises the most intriguing questions like author, creator, originality and inventiveness of AI. AI has been used in many fields like game playing, problem-solving, theorem proving, natural language processing, speech and audio recognition, computer vision,  emotion recognition and neural network. We need to understand the basic idea behind the IPR law is to protect the invention made by humans to promote innovation and hard work. By excluding AI-generated Patent or creative work, we are hampering the innovation and advancement in technology. 

Rise of artificial intelligence in recent years

Artificial Intelligence can be used in the management of IP Rights and now has become a part of our everyday life. For example, WIPO Translate and WIPO Brand Image Search which relies on AI for its functioning. At the same time, there are certain concerns towards granting copyright and patent to AI-generated works as it involves machine learning and algorithm which is opaque in itself. The rise of robot reporters such as Heligrof (Washington Post), Cyborg (Bloomberg), Bertie (Forbes Content Management System) is meant to assist reporters in doing substantive journalism, not to replace them. Dropout is one of the most widely used and efficient technologies for training the neural network. The patent of ‘dropout’ is granted in 2016 and since then there is a fear of potential patent infringement and monopolization by Google. IBM Watson is developed as a computer system to compete with TV game show Jeopardy. It is a cognitive machine which answers questions of human beings. It performs text mining and search volumes of data. Next Rembrandt, a 3D painting which produces a painting using mainly data from Rembrandt, one of the greatest Dutch painters. The painting has similar properties to the Rembrandt painting. It challenges human creativity. It won many prizes and now is used for the restoration of damaged masterpieces. E-David, a painting machine that uses visual optimization and makes paintings. It calculates brushstrokes from an input image and paints the image on a canvas. It independently take pictures and generates new creative input. 

‘Love lab’, a mathematical model which is used to predict the success of any marriage. It can predict which marriages are doomed to end with divorce. A Robot Lawyer was set up by Joshua Browder, a student of Stanford. It uses algorithms to quickly analyze a large amount of data. It was unveiled in 2016 to deal with the issue of multiple parking ticket in London. This robot helps in overturning 1,90,000 tickets with the help of deep algorithms and data given by the people. Algorithmic Contracts are contracts in which parties determine their responsibilities and obligations in the contract by using the algorithm. Currently, the common law doctrine is not applicable to such contracts, but it is currently used in areas of high speed financial products and dynamic pricing. Robots which are automatically run by using artificial intelligence in combat vehicles to target and combat threats are commonly referred to as ‘killer robots’. There is a growing fear that it will violate the Martens Clause and human rights if it is allowed to be used by the undemocratic countries. It is also argued that whoever builds the best Artificial intelligence will emerge as the victory. International regulatory framework is needed to regulate these killer robots. Recently in the late 2017, Saudi Arabia gave honorary citizenship particularly to an advanced humanoid robot Sophia. While some claim it is an advanced step in the promotion of human rights others criticized it as it is harming the public policy. The day is not too far when the concept of ‘electronic machine’ will be recognized by other countries similarly as Saudi Arabia does and AI will achieve the human intelligence and will insist for right of being recognized as ‘personhood’.There are many other such examples of Artificial Intelligence, but the moot question remains the same who owns the product generated by Artificial Intelligence. Who owns the responsibility for damages, suffering and other loss of control of these automated Artificial Intelligence? 

Failing current legal framework of copyright and patent

Ownership of Artificial Intelligence is the most grey area to delve into. There are many countries which dealt with this issue with large statutory regulations, while other countries dealt with providing few regulations and there are other countries where the statutory provisions regarding this issue are totally absent. Different legal systems have different legal interpretations regarding who is the author of a work. More specifically Article 15(1) of the Paris Text of the Berne Convention does not define who is an ‘author’ but presumes someone whose name appears in the work where the author’s identity is unknown. It establishes a general presumption about the author of the work. Countries are free to enlarge the concept of Article 15 of the Berne Convention in their respective legislation. However, the scope must be related to a natural person. 

For instance, Art. L 113-5 of French Intellectual Property Code permits the authorship of copyright to be vest in the natural person, whereas in the case of software programme copyright vested to the employer is mentioned in Art. L 113-9. Art.7 Copyright Act of Netherlands designates “authors” as the employer under whom the labour carried out in the service for creating literary, scientific and artistic work. Sec.9 of the United Kingdom, Copyright, Designs and Patent Act 1988 specifies that the author of the work is the first person for whom the copyright vests and in case of computer-generated work. Under Sec. 9(3), the author shall be taken to be the person for whom the necessary arrangements for the creation of the work is undertaken. Sec.35(2) and Sec.35(6)  of the Australian Copyright Act, where the literary, dramatic, musical or artistic work is made by the author, copyright vests to the author but if it is during the employment then the employer is the owner. Similar provisions exist in Sec. 13(1) and Sec. 13(3) of the Canadian Copyright Act. 

Sec. 201 of the U.S Copyright Act, 1976 have similar copyright laws and right vests to employers if the work is carried out under him. Art. 2(1) of EU Directives are mostly silent on the issue of authorship but the author of the computer programme is deemed to be the natural person or group of natural persons who have created the program or where the legislation of the member state permits, the legal person designated as the right holder by that legislation. Art.10(2) of the Copyright Act of Germany says that where the author name has not been mentioned, the editor of the work shall be presumed to be the author and in case if there is no mention of editor and publisher of the work is entitled to assert such right. Sec. 101 of the U.S. Copyright Act defines joint work as that the contribution of both the authors must be inseparable and interdependent. Sec. 10(1) of the U.K Copyright Act, defines joint work as the work in which the contribution of both the authors should not be distinct from each other. Art. L 113-2 of the French Intellectual Property Code recognizes collective work as the initiative by the natural or legal person who edits it, discloses it and publishes it in which the contribution of various authors have been merged. Art. L 113-5 of French Intellectual Property Code also recognizes that the collective work shall be the property of the natural or legal person, unless stated otherwise. Sec. 201 (c) of the U.S Copyright Act distinguishes between the copyright of authors of the collective work as from collective work considered as one. The notion of collective work is different in different jurisdictions.

If we analyse the Patent laws around the world particularly US Patent Act, the inventor is defined as the ‘the deviser of the invention.’ The word deviser had been interpreted slightly broader than just mere making or implementing of planning or a course of action. Inventors still have the right to be named as an inventor in US Patent law the only hurdle is that the inventor should not be the same person who applied for the patent. The four main challenges that the present Patent law all over the world facing are:

(i) The patent subject matter eligibility of AI technologies,

(ii) The patentability and inventorship of AI-generated inventions,

(iii) Liability of patent infringement by AI technologies, and 

(iv) AI’s role in the definition of “person of ordinary skill in the art (POSA)”. 

In the wake of selfie-taking monkey claim of copyright, the United States Copyright Office made it clear that the U.S. law protects only the “fruits of intellectual labour”. This case has put many repercussions of AI on US patent law. Recently in late 2017, Saudi Arabia gave honorary citizenship, particularly to an advanced humanoid robot Sophia. While some claim it is an advanced step in the promotion of human rights others criticized it as it is harming the public policy. The day is not too far when the concept of ‘electronic machine’ will be recognized by other countries similarly as Saudi Arabia does and AI will achieve human intelligence and will insist on the right of being recognized as ‘personhood’.

By analyzing the current existing legal provisions, it is clear that the laws put significant hurdles to grant copyright and patents for AI-generated works. In the EU, the process of granting copyright and patent is less stringent and recognizes the humans behind the creative process whereas in the USA impediment already exists to recognize works authored by machines.  It requires an overhaul of the existing legislation of different countries and rules and practices of Copyright and Patent Offices to make it more favourable for AI-generated works.

Conclusion and way forward

Rapid development in AI pushed the need to recognize AI-generated works by Courts and lawmakers before they confronted storms of AI-related disputes. Considering the indispensable impact and rise of AI, a potential regime should be created to protect AI as creators and inventors. The problem is more grave as neither national nor international law recognizes AI as a subject-matter. Another set of problems that needs to be addressed is to determine the terms of protection and transferability of ownership of AI-generated works. Rise of AI is challenging our preconceived notion about creativity which can no longer be done by humans only. Policymakers and judges should define the boundaries of liabilities for these AI. In the US, UK and EU, the judiciary continuously denies copyright and patents to non-human works, which is hampering the innovation process to a great extent. Many scholars have argued to recognize the non-human inventions. Researchers argued that patent rights should be granted to the inventions by artificial Intelligence because the human mind cannot achieve that much advancement in creation. 

At the same time, granting copyright and patent rights to AI will further raise issues like monopolistic behaviour which can trigger antitrust laws, fitting into the definition of POSA (Person Ordinary Skilled in Art) in IP legislative framework around the world and might emerges as potential patent infringer if AI starts writing their own codes and software. Today, AI has become more prone to be impacted by legislative interventions because of the lack of a legal framework to recognize it. This might have some negative impact and there is a need that more economic-based research should be carried out. AI may surpass all human intellect and day is not far when human-created AI will no longer require any human intervention for the grant of Patent and Copyrights.

LawSikho has created a telegram group for exchanging legal knowledge, referrals and various opportunities. You can click on this link and join: