Libertatem Magazine

Artificial Intelligence and Intellectual Property Rights: An Introspect

Contents of this Page


In our common understanding, Artificial intelligence is the ability of the computer to make decisions on its own. In today’s tech-savvy world AI is quickly gaining intense and widespread momentum for e.g. Sophia the 1st humanoid robot to receive citizenship from any country or Google’s machine learning algorithm which is capable of generating its own AIs which are advanced in many aspects than the ones created by humans, today’s AI-enabled systems have transgressed the ability to perform simple calculation and are capable of creating complex creative work such as art, poetry or well-researched article pieces such as those created by Heliograph, the AI-enabled reporter of the Washington Post. This poses a question as well as a challenge that whether the work created by these AIs is capable of affording special status under IPR just like that of humans and the impact of such determination on the jurisdiction of the law. This article mainly focuses on analyzing the interface that AI has with the specific fields of IP law and suggesting some plausible solutions for challenges and loopholes in the current legal scenario.

Intellectual Property and AI

According to John McCarthy, AI is the notion of a program, acting, and processing information, the result of which is parallel to how an intelligent man would respond with similar inputs. Today’s AI-enabled systems are capable of generating the same level of output as a human in the form of music, poetry, literature, designs, etc. and this raises a question that whether the output rendered is the result of AI’s own intelligence or of its creator. Also, copyrights to AIs have also been rejected before due to their indeterminate legal status, thus making it important for us to focus on AI’s interface with patents and copyright in particular.


The interface of AI with copyrights and patents

Copyrights are legal rights or intellectual property rights granted to a person for his/her original work or novel creations, allowing them to use such work for a definite period of time. Since copyrights are generally given for artistic and literary works, it is one of the contemporary areas of AIs applicability, so the study of it becomes relevant. Many eminent AI companies use copyrights to protect their personal data and sets of algorithms and even use digital security locks to protect the creative content of an AI yet the ambiguity on whether copyrights should be granted to an AI persists. The National Commission on New Technological Uses of copyrighted works stated in its report that “the capacity of an AI to create independent work is not practical it is theoretical” while the Office of technology suggested that “AIs can be considered as legitimate co-authors of copyrighted works”. Lovelace a major critique of the grant of copyright to an AI suggest that “machine lacks creativity because of their bounded behavior”. Many judgments by different courts also affirm this belief that the work created by an AI is not creative, for e.g. in the case of Feist publication v. Rural telephone service co. it was stated that copyright law secures only ‘the fruits of intellectual labor that are founded in the creative powers of the mind’. Also in Burrow Giles Lithographic co. V. Sarony the dichotomy between the mechanical and creative labor was addressed by the court, it was stated that a purely mechanical output by itself is not creative thus denying the grant of copyrights. While the court did provide copyright to work of an AI in Alfred Bell & co. v. Catalda fine arts but the stance taken by the court was not definite thus ambiguity still prevails in this regard.

A patent is an especial right granted for an invention that provides a novel way to its users to perform any action, it offers a new way of solving technical problems. Anyone with such rights can exclude others from selling, making, or even using the invention for a limited term. One of the requirements to attain a patent is that it should involve an inventive step, this particular point is used by the courts to exclude AIs from getting patents. In Bilski v. Kappos court denied the patent to the program by stating that what they perform is mechanical and not inventive. Also, patents are unattainable for certain facets of AI such as data accumulation, also as AI is fundamentally linked to data, patently doesn’t provide the required security as it becomes public after 20 years, thus this should be classified as trade secrets which remain for an indefinite period of time.

Thus the extent of the advancement of technology and AI particularly is outperforming its legal facets due to which there remain many loopholes and glitches which can lead to infringement and misappropriation, companies have to be alert to safeguard their IPRs in this regard.

Who is the owner of the AI-generated work and who is liable for infringement?

Take for e.g. Electronic dance music which are produced by different software mechanism, now we already know that AI can reshape and create music, it can also mimic the voice of another artist so in case if music is created by the AI software, should the copyright ownership go to the AI that has created the music or to the programmer. According to the compendium on copyrightable authorship of US copyright office practices, “US Copyright Office would confer copyrights on an original and a novel based on a condition that it must be created by a human being”.  Many court judgments have also supported this view of not granting copyrights to non-humans such as in Acohs Pty. v. Ucorp Pty. Ltd., and Bleistein v. Donaldson Lithographing Co., courts held that exclusivity of human nature is a prerequisite for copyrights and it should not be granted to a creation not created by humans. While this observation may change in the future as of now this is beneficial as first to produce an imitative work, the permission of copyright holder of original work is required which is not possible if copyright vest with the AI. Second liability upon infringement which is not possible on AI due to its intangible nature should lie upon the person who claims its ownership. This is done by Volvo who has expressly declared that they would take complete responsibility for the acts of their vehicles in autonomous mode. While this is applicable for the current scenario, we must also decide on the extent of involvement of AI and humans and cull out the ways to determine the extent of liability, apt law must also be made so that the terms of sanction can be specified for each level of liability.

Conclusion and suggestions

While the current phase of law regarding artificial intelligence is problematic where recognition of work generated by AIs is under ambiguity and there is a need for some rigid, structured, and clear legislative rules to ameliorate the problem. With the widespread usage of AI in everyday life there arises an immense need of implementing the proper guidelines urgently in the country for a starter, there must be clear-cut legislation that deals in detail with the human-machine relationship and thus draw parallels regarding ownership and liabilities of AI-generated work. Further, blockchain technology can be used to get clear information of every ensuing download and stream of a particular file to make sure that it is not infringed upon

About the Author