In Bollywood, the most popular male vocalists are playback singers whose voices can fit any film actor but who do not appear onscreen themselves. Arijit Singh is one of the few to have emerged from behind the curtain.
As of March 2025, Singh is the most followed artist on Spotify, with 138,5 million listeners, surpassing Taylor Swift’s 133 million. While Swift remains in the top 10 for streams, Singh ranks at 70. Singh’s popularity as a solo artist, however, has led to an unsettling development: companies using artificial intelligence (AI) to replicate his voice.
In 2024, Singh sued and won a landmark case, potentially establishing a legal precedent for personality rights in the age of AI. The Arijit Singh v. Codible Ventures LLP case marks the first Indian judgment addressing the misuse of generative AI tools, intellectual property (IP) and music. It also highlights the growing tension between technological innovation and personality rights, as generative AI challenges traditional norms around identity and authorship.
Arijit Singh v. Codible Ventures LLP case
So what happened? Singh alleged that Codible Ventures was using AI tools to synthesize artificial recordings of his voice, a practice known as voice cloning. It also used Singh’s likeness in its advertising, misrepresenting his endorsement of or performance at its virtual event, and created various assets bearing his name and likeness without authorization.
“What shocks the conscience of this Court is the manner in which celebrities are vulnerable to being targeted by unauthorized generative AI content”
The judges ruled that Singh’s name, voice, image, likeness, persona and other traits are protected under his personality rights and right to publicity. The court expressed particular concern about the potential for exploitation enabled by this new technology.
“What shocks the conscience of this Court is the manner in which celebrities, particularly performers such as the present Plaintiff, are vulnerable to being targeted by unauthorized generative AI content,” noted Justice R.I. Chagla of the Bombay High Court.
This ruling not only protects the rights of one of India’s most beloved singers but also serves as a critical reference point for creators worldwide navigating unauthorized exploitation of their personas in the era of AI.
India’s legal precedents for celebrity personality rights
This is not the first time an Indian court has ruled that celebrities have the right to protect various facets of their personality from unauthorized commercial exploitation, even before the advent of AI.
However, as Madhu Gadodia, Deputy Managing Partner at Naik Naik & Co., explained in a WIPO Conversation in 2024, the concept of personality rights is relatively new in India. It had to be derived from common law, copyright, trademarks and even the Advertising Standards Council of India (ASCI). The latter protects famous people from having their faces used to advertise things without authorization; this also applies to their voices, which can be just as recognizable.
A few seconds of audio is all that’s required to clone a voice with up to 95 percent accuracy
There have been several cases establishing these rights. In D.M. Entertainment (P) Ltd. v. Baby Gift House (2010), the defendants were alleged to be selling dolls that not only bore the likeness of Daler Mehndi, a popular Indian composer, lyricist and singer, but that could also sing lines from his songs. The court held that the unauthorized use of a celebrity’s persona for profit was liable for violation of the right of publicity, false endorsement and passing off. In other words, the right to publicity stems from an individual’s inherent right to choose whether to commercially exploit their identity.
In Amitabh Bachchan v. Rajat Nagi (2022), legal protection was extended to legendary actor Amitabh Bachchan against the misuse of his personality traits. It led the Delhi High Court to pass its first John Doe order, as Gadodia also noted in her 2024 WIPO talk. This was an order against the world and public at large for the protection of personality rights. Even if the defendant is unknown today, the order automatically applies to future defendants or when infringement comes to light. This includes common misuse, as well as future mediums including NFTs and the metaverse.
In rulings to protect personality rights and the right to publicity, plaintiffs must prove three key elements: their celebrity status, that they are identifiable from the defendant’s unauthorized use and that such use by the defendant is for commercial gain.
The Delhi High Court observed in Anil Kapoor v. Simply Life India (2023) that veteran actor Anil Kapoor’s right to endorsement would, in fact, be “a major source of livelihood for the celebrity.”
Therefore, any items “bearing the face or attributes of their persona on it” require their lawful authorization.
In Karan Johar v. Indian Pride Advisory Pvt. Ltd. (2024), the Bombay High Court restrained the unauthorized use of the filmmaker Karan Johar’s name and persona for commercial gain.
Arijit Singh’s personality-rights victory against AI
In the case of Arijit Singh, the Bombay High Court also ruled that the singer’s personality attributes, including his name, voice, photograph and likeness, were protectable and that the unauthorized creation of merchandise, domains and digital assets was illegal.
The court further found the following protectable facets of Singh’s personality rights and right to publicity: his voice, vocal style, vocal technique, vocal arrangements and interpretations, his mannerisms and manner of singing, and even his signature.
Moreover, the court found that the use of AI tools to recreate Singh’s voice and likeness – apart from violating his exclusive right to commercially exploit his personality – could potentially affect his career if used for defamatory or nefarious purposes.
“This form of technological exploitation infringes upon the individual’s right to control and protect their own likeness and voice”
Singh had successfully defended his personality rights by securing an ad-interim order from the Bombay High Court, preventing several entities, including AI tool platforms, from commercially exploiting them.

How AI voice cloning works
Real-voice cloning or retrieval-based voice conversion (RVC) is an open-source AI algorithm that allows users to clone voices from pre-existing audio samples. With easy-to-use tools such as Jammable or TopMediai, a few seconds of audio is all that’s required to clone a voice with up to 95 percent accuracy. These voice-cloning tools pose a substantial risk to IP right holders and the public at large through potential fraud, such as scammers cloning the voice of a friend or loved one and using it to request money.
AI tools need to be trained on data in order to generate new works, which feature attributes of the training dataset. In Arijit Singh v. Codible Ventures LLP, it was alleged that a dataset comprising 456 songs from Singh’s repertoire was uploaded to AI tools without authorization. These tools enabled users to convert any text, speech, voice recording or audio file to Singh’s AI voice version.
The challenge of AI voice cloning exists at the intersection of personality rights and copyright protection. For singers such as Singh, voice is both a personal attribute and the medium for creating copyrightable works. When AI tools are trained on an artist’s songs without permission, they simultaneously infringe on the artist’s copyright while extracting and replicating their distinctive vocal characteristics.
Laws need to be applicable not only to celebrities but to every individual, as each person has a right to protect their personality and their privacy
In Singh’s case, the defendants also attracted visitors to their websites and/or AI tools by capitalizing on Singh’s popularity and reputation, thereby subjecting Singh’s personality rights to potential abuse. The defendants also emboldened internet users to create counterfeit sound recordings and videos that misused Singh’s character and identity.
The Court held that “making AI tools available that enable the conversion of any voice into that of a celebrity without his/her permission constitutes a violation of the celebrity’s personality rights.” They also found a celebrity’s voice to be “a key component of their personal identity and public persona.”
“This form of technological exploitation not only infringes upon the individual’s right to control and protect their own likeness and voice but also undermines their ability to prevent commercial and deceptive uses of their identity,” noted Justice R.I. Chagla of the Bombay High Court.
The Court observed that allowing the defendants to continue using Singh’s name, voice, likeness, and so on, in the form of AI content without consent would not only risk severe economic harm to the singer’s life and career but also leave room for unscrupulous individuals to misuse such tools for nefarious purposes.
What stands out is how the courts are finding ways to apply existing legal frameworks (personality rights, copyright) to these new challenges rather than dismissing them as too novel to address. While technology evolves apace, fundamental principles about personal identity and the right to control one’s own expression remain important.
About the author
Prof. (Adv.) Dipak G. Parmar is an IP Attorney, mediator and arbitrator and founder of Cyber-IPR in Mumbai, India. He serves on the Advisory Board of the Centre for Development of Intellectual Property and Research (CDIPR) and the Council of EU Chambers of Commerce in India, among other memberships.
‘In the Courts’ articles typically report on current court cases and rulings and are circulated in a timely manner for discussion and comment.