Rethinking Authorship: Legal Dilemmas in Copyright for Ai-Generated Content

Rethinking Authorship: Legal Dilemmas in Copyright for Ai-Generated Content

This piece has been authored by Kumari Bhargavi (4th Year BA. LLB (Hons) student at Symbiosis Law School, Pune)

Keywords: Artificial Intelligence, Legal Entity, Copyright Laws, Authorship Rights, Human Creativity     

INTRODUCTION

The evolution of technology in the modern era plays a vital role in the development of a country and its individuals considering development under the ambit of education, transportation, medical science, research, as well as the entertainment sector. Owing to the introduction of AI generated content, it is not an alien concept to the field of intellectual property rights, since it plays a significant role in generation of art and literary, alongside providing information as desired by the users.

Under the framework of provisions related to copyright laws, protection is provided to human generated works, as well as authorship of work to human who provide creative input to obtain machine generated output. However, a legal dilemma arises when it comes to providing protection to the work solely created by AI devoid of human intervention.

Against this backdrop, the author seeks to the analyse the lacunae created by AI generated content in area of copyright law pertaining to authorship rights. This article aims to analyse whether authorship rights should be granted to AI or AI- generated works should fall under public domain, considering the legal position of few developing countries on these issues.

EXPLORING THE INTERSECTION OF ARTIFICIAL INTELLIGENCE AND COPYRIGHT LAW THROUGH LEGAL PERSPECTIVE

In contemporary times, AI has gained significant importance in various fields that aid in the sustainability of humankind, as it eliminates errors and increases efficiency by automating many activities and reducing human involvement. In 1956, John McCarthy coined the term “Artificial Intelligence.”[i] However, no legal statute has yet provided a definition of AI.  In common parlance, AI often refers to a machine’s capacity for cognitive functions including thought, perception, acquisition of knowledge, and decision-making processes. While taking into consideration the Indian perspective, Section 2(d) of the Copyright Act of 1957[ii] defines author as someone who creates copyrightable works that are inclusive of creative works like music, drama, literacy, or any software generated work.

AI, as described by WIPO, is divided into two subsets: ‘machine learning AI’ and ‘deep learning AI.’ While the former uses pre-programmed input to generate output, the latter, an extension of machine learning AI, uses artificial neural networks, enabling the classification and grouping of unlabelled data. Analysing the relation between AI and copyright laws, the “issue of granting authorship to AI” arises, since AI produces creative works that are either “AI-generated”, i.e., without human intervention, or “AI- assisted”, that refers to works created with human intervention. Authorship rights could be granted to the programmer or user for AI-assisted work. However, for AI-generated work, issues arise concerning authorship rights and whether AI should be considered a legal entity to determine the goodwill of the works and the associated rights and liabilities. AI-generated work might contain unanticipated data or have perilous impacts, thereby creating loopholes.

ARTIFICIAL INTELLIGENCE AND THE COPYRIGHT DILEMMA

Prof. Stephen Hawking stated that the extensive development of AI might lead to the extinction of the human race. If legislations recognize AI as authors, then human creativity and software creativity would be on the same pedestal and in the long run it might give rise to the decline of human creativity. Further, sometimes AI- generated work might give unwanted output that might be provoking in nature or be obscene, which would be detrimental for the younger generations. In such cases, even if authorship rights were granted to AI, it would be difficult to determine civil and criminal liabilities due to its lack of personhood.

The “Copyright Act, 1957” provides moral rights to the authors, divided into ‘right to paternity’ and ‘right to integrity’  under Section 57[iii], to claim rights against copyright infringement, from the case of Amar Nath Seghal v Union of India[iv], it was observed that moral rights are influenced by the thoughts and sentiments of the human author, and AI, not being human, is not intended to be covered by these rights.

Another argument in favour of not granting authorship right to AI is that if same input is provided to the AI each time to obtain output, the output so generated would be similar in nature, thereby lacking creativity. There are views that works produced by artificial intelligence ought to stay in the public domain for its use by public for free, however, the stated purpose of copyright laws is to motivate the authors to continue developing novel works, for the benefit of society by giving them incentives in the form of financial and moral right. If users start commercializing the output so generated in several ways than it would cause loss to the company that invents AI software, thereby making them less inclined to invest in AI, and hence it acts as a primary argument to not grant authorship rights to AI.  Furthermore, the concept of fair dealing as mentioned under Section 52[v] should also be considered. Section 2(z) defines “work of joint authorship” stating that nature of contributions of multiple authors should not be different[vi], hence, considering human author and AI as joint authors could give rise to grey areas under this definition. 

The European Parliament suggested considering “autonomous robots” as legal entities, to grant them protection and rights under the copyright laws. In 2017, the AI humanoid robot, Sophia, was granted citizenship rights by Saudi Arabia. The rights enjoyed by the robot were more than those enjoyed by the females of that country. Further, with the increasing use of AI programmes like ChatGPT, the use of academic malpractice has increased.  These instances give rise to the idea that granting authorship rights for AI- generated work, and considering AI as a legal entity is the need of the hour, however, the rights granted to AI should not give it preference over human creativity. International treaties often outline the minimal universal norms that needs to be followed. Although the nations must adhere by them, they are also free to provide greater protection than what is allowed under the treaties[vii].

CAN THE CURRENT LEGAL FRAMEWORK ADEQUATELY SAFEGUARD AI-GENERATED WORKS?

The author conducts a comparative analysis with a focus on contemporary statutes of Japan, China, Germany, the United Kingdom, and India to examine copyright protection for works generated by A.I. with human intervention.

Under the Indian Copyright law, it is evident from Section 2(d)(vi) that ambit of “author” is not limited to living entities or human, since authorship rights can also be granted to works that are software generated[viii]. This could be substantiated by taking into account, “RAGHAV”, an AI painting app, that has been authorized in 2021 as a co-author of the copyrighted work “Suryast”. In the United Kingdom, copyright protection extends to AI-generated works under Section 178 of the “Copyright, Designs, and Patents Act of 1988” that provides protection to AI-generated works.[ix] Furthermore, under Section 9(3) it has been stated that the individual making essential arrangements for the software-generated work would have the ownership rights of the said work[x]. This framework highlights the UK’s progressive approach in recognizing and protecting AI-generated intellectual property.

Countries like Japan, China, and Germany do not consider AI as authors because AI-generated outputs lack originality and minimal creativity, which are the parameters for granting copyright protection. However, in the judgement of Shenzhen Tencent v Shanghai Yingzun[xi] the Chinese Court granted protection to works that are AI- assisted, reflecting upon how the approach of countries with respect to granting authorship rights to AI-generated works are evolving.

CONCLUSION

Artificial Intelligence will become increasingly pervasive in many facets of our everyday lives. Its applications must be governed by legislation. AI will continue to play a crucial role in the world of intellectual property rights, especially in copyright. The border between human and machine creativity has become blurry as AI has developed, giving rise to a complicated discussion over the attribution of copyright laws and how the law should be applied to AI-generated works. Ineffective management of authorship rights for AI-generated works might result in an accumulation of authority and creative monopolies. It is comprehended that a comprehensive legal framework should be adopted to regulate the authorship rights of AI- generated works, further, human ingenuity should be valued above artificial creativity. Therefore, it is imperative to adopt a balanced strategy.


[i] Fredy Sánchez Merino, ‘Artificial Intelligence and a New Cornerstone for Authorship’ (2018) in WIPO-WTO Colloquium Papers <https://www.wto.org/english/tratop_e/trips_e/colloquium_papers_e/2018/2018_complete_file_e.pdf > accessed 08 October 2024

[ii] Copyright Act 1957, s 2(d)

[iii] Copyright Act 1957, s 5

[iv] Amar Nath Seghal v Union of India, [2005] 30 PTC 253 (Del)

[v] Copyright Act 1957, s 52

[vi] Copyright Act 1957, s 2(z)

[vii] Ritika HR, ‘AI-generated works and Copyright: Navigating the complexities and intersection of AI-IPR in the Digital Age’ (2023) 3(1) DSNLU Journal of Science, Technology and Law < https://dsnlu.ac.in/storage/2023/09/DSNLU-J.-SCI.-TECH.-L.-2023.pdf> accessed 03 October 2024.

[viii] Copyright Act 1957, s 2(d)(vi)

[ix] UK Copyright, Designs and Patents Act 1988, s 178

[x] Copyright, Designs and Patents Act 1988, s 9(3)

[xi] Shenzhen Tencent v Shanghai Yinxun, [2019] Yue 0305 Min Chu 14010

Leave a Reply

Your email address will not be published. Required fields are marked *