[A big thanks to Praharsh for his inputs on the post.]
On July 26, 2024, the Bombay High Court (BomHC) delivered a significant order in the case of Arijit Singh v. Codible Ventures LLP, marking a big step in the protection of artists’ personality rights against the unauthorised use of their voices by artificial intelligence (AI) tools. In this case, noted singer and composer Arijit Singh requested legal protection against the unauthorised commercial use of his personal attributes, such as his voice, name, signature, image and mannerisms, to sell various articles online. Two side notes before getting into the meat of the order: First, it is interesting to see that though the plaintiff raised the argument of the infringement of moral rights, the court hasn’t addressed this. Second, in different places, the Court has used the language “Personality rights and right to publicity”. It is unclear why the Court has separated these two when they seem to have addressed ‘personality rights’ as the right to publicity. Conceptually, as there isn’t a ‘personality right’ per se; the two origins for ‘personality rights’ have either been privacy rights-based, or publicity rights-based.
Coming back to the ruling: This ruling seems to be the first of its kind in India to directly address the legal complexities surrounding AI-driven voice cloning, setting a precedent for how the law might adapt to rapid technological advancements. The BomHC’s order underscores the growing need to reconcile intellectual property rights with the evolving capabilities of AI. This order is noteworthy as it sets in motion a series of questions about how this technological advancement will be regulated to safeguard artists’ rights in an age where technology can easily replicate and exploit their work and how the system of IPR can work in tandem with technological growth. This post (Part I) aims to analyse the BomHC’s findings and observations, while Part II (here) will raise questions about the implications of this order for future disputes.
Copycat Cadence: Understanding What ‘Shocked the Conscience’ of the BomHC
In the order, Justice Riyaz Chagla noted that the case ‘shocked the conscience’ of the BomHC, highlighting the vulnerability of celebrities, particularly performers, to being exploited by unauthorised AI content creators. Acknowledging the substantial goodwill and reputation that Arijit Singh has cultivated, the BomHC noted with concern that the defendants were leveraging the plaintiff’s popularity to attract visitors and drive traffic to their websites and AI platforms. The court stated that this exploitation not only risks infringing upon Arijit Singh’s personality rights but also “could potentially jeapordize the Plaintiff’s career/livelihood”. The bench emphasised that while freedom of speech and expression permits critique and commentary, it does not authorise the commercial exploitation of a celebrity’s persona for financial gain.
Upon examining the Plaintiffs’ arguments and drawing on precedents such as Amitabh Bachchan v. Rajat Nagi (discussed here), Anil Kapoor v. Simply Life India (discussed here), D.M. Entertainment v. Baby Gift House (discussed here), Applause Entertainment v. Meta Platforms (discussed here), and Karan Johar v. Indian Pride Advisory, the BomHC noted how the courts have consistently recognised and enforced personality rights. The established legal principle in these cases is that proving a plaintiff’s celebrity status is merely the initial step in safeguarding personality rights and the right to publicity.
In the present matter, BomHC observed prima facie that the Defendants are allegedly engaging in unauthorised use of the Plaintiff’s personality traits, including their name, image, and likeness. This use is such that the Plaintiff can be distinctly identified. It further noted that this exploitation was being conducted without the Plaintiff’s consent and was pursued for both commercial and personal benefit. BomHC highlighted that these activities are not only unauthorised but also infringe upon the Plaintiff’s established personality rights. A table of all the various allegations of activities infringing the singer’s rights is provided below for a better understanding of the “range” of activities that can fall under the ambit of protection under this order:
| AI Platforms and their Operators (Defs 1-8): These defendants are AI platforms or their owners/operators that use advanced algorithms to create audio and visual content mimicking the Plaintiff’s name, voice, mannerisms, and likeness. They exploit the Plaintiff’s personality traits to capitalise on their goodwill and reputation. |
| AI Music Creation Website (Def 4): They run an AI platform for creating music using AI models of famous singers. The website includes AI voice models of the Plaintiff, allowing users to input YouTube links or audio files to produce outputs in the Plaintiff’s AI voice. This usage of the Plaintiff’s likeness is unauthorised and has been documented in videos and screenshots. |
| AI Platforms (Def 6): They operate two AI platforms, one of which has a blog and tools for converting text/speech into the Plaintiff’s AI voice. The site provides step-by-step instructions for unauthorised conversion. On the other there are guides for converting text or speech into the Plaintiff’s voice, with links to additional AI tools like voxbox and magic mic. Documentation includes screenshots and screen recordings from these sites. |
| Music Production Company (Def 7): They produce music and conduct courses using AI tools, Their YouTube channel “Basslila” uploaded a Video which unauthorizedly uses deepfake technology to convert text or speech into the Plaintiff’s AI voice, exploiting their voice, mannerisms, and likeness for commercial gain. |
| AI Voice Cloning Platform (Def 3): Defendant No. 3 operates a platform that converts speech or voice recordings into Plaintiff’s voice using Real Voice Cloning (RVC) technology. They have unauthorizedly uploaded 456 songs from the Plaintiff’s repertoire, enabling users to convert text or audio files into the Plaintiff’s AI voice. Additionally, one of the founders of Defendant No. 2 has promoted these activities on social media, providing guidelines for unauthorised voice conversion using Defendant No. 3’s platform. |
BomHC issued an ex-parte ad-interim order that places significant restrictions on several third parties regarding the use of Arijit Singh’s personal attributes. Specifically, it prohibited these parties from using his name, voice, vocal style and techniques, mannerisms, photographs, images, signature, persona, or any other aspect of his personality for any purpose—commercial or personal—without his explicit consent.
The injunction protects Arijit Singh’s personality rights across all media, including physical, digital, and the Metaverse. It covers the use of any technology or application exploiting his traits, such as online platforms, publications, ads, merchandise, and domain names, as well as advanced AI technologies like voice models, voice conversion, synthesised voices, digital avatars, caricatures, deepfakes, face morphing, and GIFs. This wraps up the discussion on the BomHC’s analysis of the issue of AI cloning of singers’ voices. Part II of this post (here) discusses the international situation surrounding AI voice cloning and raises some questions for consideration for similar cases in the future.