Introduction
In 2023, the release track ‘Heart on My Sleeve’ by a famous TikTok user named Ghostwriter977 took the music industry by storm as the composition which seemed like a collaboration between Drake and The Weeknd was actually generated by artificial intelligence and made to sound like the two artists.
This incident highlighted the growing potential of artificial intelligence (AI) in creative space and raised significant legal questions about copyright and the rights of artists.
In the Indian context, this incident has profound implications for the music and entertainment industry. This article explores two pressing issues exploring the same:
- Can AI-generated work be copyrighted in India?
- Does AI recreated the style of an artist amounts to inspiration or infringement?
Can AI-generated work be copyrighted in India?
The Indian Copyright Act, of 1957 (Copyright Act) safeguards original literary, dramatic, musical, and artistic works. A fundamental requirement for copyright protection in India is that a work must exhibit “a modicum of creativity” as emphasised by the landmark judgment in Eastern Book Company V. D.B Modak[1]. For a work to be considered original, it must demonstrate intellectual effort and substantial variation from preexisting content.
Artificial intelligence tools generate output by processing vast datasets that include content from various sources. While these outputs may not directly replicate these sources, they raise significant questions about whether they possess sufficient creativity to warrant copyright protection. Although generative AI relies on pre-existing works, its output does not merely elaborate or rearrange existing material, and if the work reflects discernible intellectual effort, creativity and distinctness, it could meet the modicum of creativity threshold as required for copyright protection.
Ownership of AI-Generated Works
In some jurisdictions, such as Canada, AI has been recognised as a co-author, while others, like the United States, maintain that copyright must be claimed by a human creator only. In India, these issues remain unresolved with courts and policymakers yet to establish a clear stance.
Section 2(d)(vi) of the Copyright Act defines “author” in relation to computer-generated work as the “person who causes the work to be created,” which inherently implies human authorship. However, ambiguity persists in determining whether it is the programmer who develops the AI system or the user interacting with the AI who should be considered the rightful author. Even if AI were granted legal personality, it would still be ineligible to assert copyright as Indian copyright law currently mandates that the creator must be a human being.
Despite this AI generated works require protection to ensure that copyright law continues to foster creativity. A prominent example of this challenge is the case of DABUS (Device for the Autonomous Bootstrapping of Unified Sentience)[2] an AI system that autonomously developed two inventions. While some jurisdictions such as South Africa and the Federal court of Austria have recognized AI as an inventor in patent applications this perspective remains rare globally.
The solution to this problem could be the recognition of co-authorship, where both human creators and AI contribute to a work. If AI were recognized as a co-author, determining copyright protection periods would require careful consideration due to AI’s “immortality.” To address this, copyright laws should introduce a licensing system specifically designed for works created by two authors—one human and one AI. This approach would help strike a balance between rewarding human creators and recognizing the role of AI in the creative process, ensuring fair compensation and protection for all contributors.
Does AI recreate the style of an artist amounts to inspiration or infringement?
An artist’s voice is not inherently protected under the copyright laws. However, the voice of the singer embodied in the sound recording or music, including AI-generated cloned work, can be protected, as fixation is a sine qua non of copyright protection.
AI-generated music raises significant legal questions, particularly around the issues of inspiration versus infringement. In the music world infringement claims often arise not from direct copying of musical works or sound recordings but from creating songs that carry a similar feel to another artist’s style and voice. This presents ample room for AI music creators to avoid direct copyright infringement by relying on independent fixation where AI generated music independently without reproducing existing sound recordings. However, when an AI system is trained on a specific artist’s work the resulting output can carry substantial stylistic similarities and voices even if it doesn’t directly reproduce[3] any sound records. This raises legal questions about the potential for copyright violations.
In consideration of that, an artist’s voice is not inherently protected under copyright laws. However, the singer’s performance embodied in a sound recording including AI-generated cloned works can be protected. In Experience Hendrix LLC v. Purple Haze Records Ltd,[4] it was held that a performer’s voice could be safeguarded through the performance though the voice in isolation is generally not protected.
Moreover Section 57 of the Copyright Act in India grants authors certain “Moral Rights” such as the right to protect the integrity of their work. For instance, in the case of Genda Phool, a remake of an old Bengali song originally sung by Ratan Khar, allegations arose that the performance and moral rights of the original artist were disregarded.
The landmark case of R.G. Anand v. Delux Films[5] established that copyright infringement is determined by whether an average listener, having heard both works, gets an unmistakable impression that the subsequent work is a copy of the original. Music as a form of art is classified into two types of copyright protection firstly the musical work which includes rhythm, compositions, and lyrics, and secondly, the sound recording itself or the master recording generated music must be analysed under both of these copyright regimes.
Regarding musical works, AI applications typically create music by mining the performative nuances from sound recordings rather than using the musical work themselves. This means that substantial portions of a song’s composition may not manifest in the final AI- generated output[6]. However, cases must be assessed individually as some portions of musical works may appear in AI generated songs, leading to potential copyright infringement if recognizable to an average listener.
The ‘Fair Dealing’ exception to copyright infringement allows for the bona fide use of copyrighted works without prior permission. In cases where AI-generated music fails the average listener’s test it is crucial to examine whether it could fall under this defence. The Civic Chandran v. Ammini Amma[7] matte laid down factors for determining fair dealing, such as the quality of the matter taken, the purpose for which it is used, and the likelihood of commonalities between the works. In light of AI-generated music, courts have increasingly placed importance on the transformative nature of the work whether it differs in character, expression, and meaning from the original work. Given that AI-generated music can easily be classified as transformative, this emphasis on the transformative factor might provide an opportunity for AI-based musicians to sidestep infringement claims even when an existing person’s music heavily influences their creations.
AI-generated music introduces both exciting creative possibilities and complex legal challenges. Musicians wary of potential copyright infringements view AI with suspicion particularly in India where copyright protection for music has historically been weak. Some solutions, such as new licensing and royalty-sharing arrangements, have been proposed to address this issue whereby AI music creators would need to obtain licenses from musicians to use their work to train AI systems and share profits from AI-generated songs. While the feasibility of these solutions remains to be tested, global trends suggest that Indian courts will soon face infringement claims against AI-generated music. In adjudicating such cases, courts must strike a balance between fostering creativity and protecting the rights of musicians.
Conclusion
AI-generated music challenges traditional copyright frameworks blending innovation with complex legal issues. Currently, Indian copyright law focused on human authorship requires reform to address co-authorship models and licensing systems for AI contributions. Courts must carefully assess the line between inspiration and infringement while accounting for the transformative nature of AI works.
Balancing between fostering innovation and protecting artists’ rights is essential. Proactive reforms will ensure that India’s creative industries adapt to technological advances while safeguarding the rights of creators.
[1] 1 SCC 1; 2002 PTC 641.
[2] DABUS: South Africa issues first-ever patent with AI inventor | Managing Intellectual Property
[3] AI & The Music Industry; Copyright Infringement & Creative Overlaps – Copyright – India
[4] [2003] EWHC 1315 (Ch) (Eng.).
[5] 1978 4 SCC 118
[6] Artificial intelligence in music: recent trends and challenges | Neural Computing and Applications
[7] 1996 SCC ONLINE KER 63
