The music industry is not an exception to the surge of innovation and transformation brought about by the rise of generative artificial intelligence (AI). With the help of this ground-breaking technology, it is now possible to create music entirely generated by algorithms, upending preconceived ideas about artistic originality and creativity. In light of the unprecedented surge in viewership, with millions of views amassed within a mere fortnight of its debut, it behooves us to underscore the fact that the mastermind behind the composition of the opus entitled "Heart on my Sleeve" is none other than the esteemed individuals, Drake and The Weeknd, despite their absence of ownership stake or entitlement to any pecuniary proceeds stemming from the aforementioned musical creation. By all accounts, this musical piece is indisputably an embodiment of their artistic prowess, albeit devoid of any legal prerogative they may assert over it.
In the context of India, this article aims to scrutinize the definition of the author of a musical work and the accompanying rights of such work and engage in a comprehensive exploration of the potential arguments regarding ownership of the work put forth by both AI and the artist who may have been impersonated.
The Enigma of Ownership: Decoding Rights in AI-Generated Musical Works
The scope of the Copyright Act of 1957 (herein referred to as "the Act") extends to protect original literary, dramatic, musical and artistic works and cinematograph films and sound recordings from unauthorized uses. A "musical work" as per the Act refers to a work consisting of music and includes any graphical notation of such work but does not include any words or action intended to be sung, spoken or performed with the music.
Although the Act protects original artistic, musical, and literary works, it does not provide a specific definition for the term 'song'. Consequently, the Act protects the distinct elements comprising a song as a collective entity. The multifaceted rights encompassed in a singular song's creation can be elucidated through the following example. Let us consider a scenario wherein an individual, X, renders a vocal performance of a song, the lyrics of which are authored by Y, while the musical composition is crafted by Z, all for the production of a motion picture helmed by a party referred to as A. X, as the singer, assumes the role of a performer in accordance with the provisions outlined in Section 38 of the Act. Y's creative output, comprising the lyrical composition, garners recognition as an original literary work, thereby receiving protection under Section 13(1)(a) of the Act. Z, the composer, is accorded the status of an author concerning the musical composition pursuant to Section 13(1)(a) of the Act. Now, the term "author" refers to the individual who causes the creation of a computer-generated literary, dramatic, musical, or artistic work. However, concerning AI authorship, it should be noted that neither the Copyright Act nor the Copyright Rules provide a specific definition for the term "person" under Section 2(d)(vi) of the Act. The General Clauses Act 1897, on the other hand, defines "person" in a comprehensive manner, encompassing both artificial and juristic persons.
In the context of a musical work, an additional complexity arises. While the producer is a legitimate individual, the AI assumes the role of an actual singer without obtaining consent from the real singer whose voice has been incorporated into the work. This places an added burden on the legal framework. In this case, there are arguments presented by both sides.
The Debate over Copyright Violation
The specific question of whether the producer/composer of the piece retains ownership, whether the AI platform, which has undergone training, can claim ownership, or if the original artist, whose voice has been imitated, retains the rightful ownership. Under specific circumstances, the ascription of authorship to the human creator remains readily identifiable when employing computer systems or software to replicate their artistic ingenuity and produce musical compositions. However, as the utilization of software or AI extends beyond a mere facilitative tool to enhance the creative features of human authors with minimal human input, the determination of authorship becomes increasingly intricate. India is among the jurisdictions that have addressed the issue of authorship for computer-generated works, aligning itself with the United Kingdom. While this provision may appear to resolve the ambiguity surrounding authorship in AI-generated works, a significant obstacle still remains.
From the producers' perspective, the voice of an artist can be understood as the result of training a Recurrent Neural Network within the AI platform using a comprehensive set of elements created by that artist. With each input segment, the model advances in its ability to emulate the sound and adapt uniquely. Through continuous learning, the technology extracts the essence and characteristics of the music in its entirety. Subsequently, the AI model is developed, possessing the capacity to compose entirely distinct, innovative, and original musical pieces. Upon receiving the appropriate lyrics, this AI platform produces an outcome that emulates the original singer's performance. The resulting song possesses unique lyrics and music, completely distinct from the original composition, devoid of any replicated components. Therefore, why should copyright infringement come into the picture?
From the side of the artist, the argument would be twofold - Firstly, the work being a direct copyright violation which fails the four-judge test of Fair Use, and secondly, the argument of publicity.
In Rogers v. Koons, the court's opinion can be interpreted as the key was to analyze the impact on the potential market as a crucial aspect of the fair use analysis. It involves evaluating whether a copyrighted work hampers the copyright owner's ability to generate income or undermines the emergence of a new or prospective market for the work. Engaging in actions which deprive a copyright owner of their rightful income is likely to be considered as an effect of copyright violation which does not come under fair use exceptions.
However, a compelling argument in favour of this notion is exercising the right to publicity by any musical artist. Unlike the United States of America, where the Lanham Act of 1946 and state laws about publicity exist, India lacks an independent statute governing these rights. Instead, the deprivation of an individual's right to publicity is viewed as a consequence of the Right to Privacy thereby infringing upon Articles 19 and 21 of the Constitution. In India the Delhi High Court established that an individual's voice associated with a musical work can confer the right of publicity upon the author. Sections 38 and 57 of the Act consider using a famous person's identity for advertising without their consent. The issue lies not in disallowing commercial gains from such usage but rather in ensuring that the famous person retains the power and authority to control the circumstances surrounding the use of their identity. This violation of the right to publicity is enshrined within these sections. Consequently, the artist would ideally only need to initiate legal proceedings stating that the (in this case, AI-generated) work has tarnished their reputation and goodwill through unauthorized use of their name, as declared by the Madras High Court in the case of Mr. Shivaji Rao Gaikwad v. M/S. Varsha Productions.
Balancing Copyright and AI
Arguments from both producers, emphasizing the innovative nature of AI-generated compositions, and artists, asserting copyright violations and the right to publicity, create a multifaceted debate. In conclusion, as technology continues to evolve and AI becomes increasingly involved in the creation of artistic works, it is imperative to introduce amendments to the copyright act that address the unique challenges posed by changing technologies. Striking the right balance between artists' rights to protect their creations and their freedom to innovate has always been at the core of copyright law. However, with AI's growing role, it is essential to navigate these complexities carefully. The Indian courts' interpretation of future cases will play a pivotal role in determining whether AI assumes a significant role or if copyright law remains steadfast in protecting the original author. Ultimately, a thoughtful and forward-looking approach to copyright law will ensure that artists' rights are safeguarded while embracing the potential of AI in the realm of artistic expression. Striking a delicate balance between innovation and artistic protection will be crucial as India navigates the copyright landscape in the realm of AI-generated music.
Aryan R. Nair, AI-Generated Musical Works in India: Copyright Conundrums and Legal Perspectives, Digital Law Asia (Aug. 2, 2023), https://digital.law.nycu.edu.tw/blog-post/hqlfit/.