Echoes Without Consent: AI Sound-alike Performances, Moral Rights, and India’s Legal Vacuum

India’s entertainment sector has witnessed an incredible transformation under the stewardship of Artificial Intelligence (AI). On one hand, we might see generations of soundalikes for the Kishore Kumar voice in advertisements, the tribute bands generating talents of their own, and the other lookalike manifestations through deepfake visuals. Yet, these advances come with their fair share of legal and ethical issues. First came the controversy in 2023 with a viral deepfake of Rashmika Mandanna, and then came the election chaos of 2024 induced by the deepfakes of Ranveer Singh and Aamir Khan. Just like in these cases mentioned, unauthorized copies of artists desecrate the very sanctity of the artist and his/her work and destroy the public trust.

Moral rights are granted to authors to protect their works under Section 57 of the Copyright Act, 1957, but it remains whether this protection would extend to protect voices or personae of actors as well? Another question that arises in this context is whether these are adequate protections against the onset of AI, or would personality rights be a much stronger defence?

This article covers moral rights, vocal style protection, lookalike performance issues, comparative insights, and policy recommendations, addressing a critical gap in India’s entertainment law.

Moral Rights in India: Scope and Gaps

The Indian Copyright Act, 1957 has provisions for moral rights in Section 57, which gives authors the right of paternity (to claim authorship of the work) and the right of integrity (to prevent distortion or mutilation of the work that may harm the reputation). These moral rights are perpetual and inalienable, like the Berne Convention, and are enforced by authors or their heirs. Section 38 of the Copyright Act provides additional protection for the right of performers against unauthorized fixation or broadcasting of performances such as a singer’s live show. Unfortunately, such provisions do not help with sound-alike covers of voices such as Kishore Kumar's case. The case of Amarnath Sehgal v. Union of India has gone as far as the Delhi High Court ordering the compensation of a work (a mural made by a sculptor) by holding that the moral rights of the sculptor had been violated, but it was limited to the protection of physical works, not a style of voice or persona.

Section 57 applies to works, means literary, dramatic, musical or artistic works in terms of the definition stated in Section 2(y). This indicates that performers are unprotected against AI generations because their vocal timbre can be replicated without having to effectively copy or reproduce the copyright songs. There is no Indian case law or authority squarely on point when it comes to whether AI imitations infringe moral rights creating a legal blackhole. Section 38 does protect performances but not the identity or persona of the performer which limits its value against 'sound alike' impersonators. This gap has to date, never been evaluated by a court in India, which also coupled with Brennan's comments on AI in Entertainmentclearly highlights the area requiring legislative response to clarify protections for performers in an AI landscape.

Can Vocal Styles Be Protected?

Does a performer’s voice, such as Lata Mangeshkar’s, qualify as a “work” under Copyright Act, 1957?

Section 2(y) defines “work” as literary, dramatic, musical, or artistic creations but does not include vocal styles or useful gestures. As a result, the moral rights under Section 57, which protect authorship and integrity, do not apply to voices. Therefore, cover versions whether sung by humans or AI (Resemble AI, Audimee, etc.), follow the sound alike aspect and escape control unless they copy something specifically copyrighted. Even if personality rights may provide an inkling of hope. In Anil Kapoor v. Simply Life India, the Delhi High Court restrained the use of Kapoor’s likeness and voice by the trademark defendant in AI use without authorization, based on privacy and passing off but not moral rights.

While redress may be considered by the performers for voice appropriation, the unwritten nature of personality rights leads to inconsistencies, especially in cases controversially rejected by courts involving unknown or celebrity performers, or in cases involving right holders of performers who died some time ago, such as in the case of Kishore Kumar. Any such deepfake, making use of AI for precisely mimicking a voice and its nuances for profits, can cause reputational harm to actual singers and also the commercial opportunities in existence for those actual singers, rendering the performers freely open to exploitation without any specific legislative protection.

Fair Dealing and Sound-alike Performances

Section 52 of the Copyright Act, allows fair dealing for certain purposes, such as criticism, review, or parody. The growth of AI-generated sound-alike performances like Instagram creators recreating parts of songs using the voice of, say, Mohammad Rafi and using it for commercial purposes creates a new problem. In Super Cassettes Industries Ltd. v. Chintamani Rao, the Court accepted parody for fair dealing purposes, where it did not affect the market for the original work. But AI sound-alike imitation has the potential to adversely impact a performer’s market value or reputation, so it is difficult to fathom this would qualify as fair dealing.

Lookalike Performances: Legal and Ethical Issue

Tribute bands imitating the performance of artists like Kishore Kumar can defensibly avoid any copyright infringement but invoke ethical concerns regarding appropriate consent and reputational damage. In Amitabh Bachchan v. Rajat Negi , the courts have granted injunctions upon unauthorized use of likeness relying on personality rights. In Jackie Shroff v. The Peppy Store, the court restrained unauthorized use of Shroff’s image and voice in AI-generated content, also focusing upon personality rights.

There are no generally accepted codes of practice; therefore, legal protections for likeness are applied inconsistently. The ethical risks increase in contemporary digital societies as the AI associated with deepfakes erodes artistic legacies. A flowchart titled Step-by-Step Guide to Challenging Unauthorized Deepfakes in India are as follows: (i) file complaint with nearest Cyber Crime Cell/Police Station as per IT Act Section 66E, (ii) seek injunctive relief with Copyright Act section 55(1) (iii) claim damages for breach of personality right (actual or punitive).

Comparative Insights: US and EU models

In the United States, under a right of publicity law, performers’ voices are protected, as shown in Midler v. Ford Motor Co. , where Bette Midler received damages for unauthorized use of a sound-alike. The European Union General Data Protection Regulation (GDPR)considers deepfakes as violations of personal data and Denmark’s copyright laws in 2025allow for voice ownership rights, which is also useful. These mechanisms differ from Copyright Act, 1957, which makes moral rights applicable especially for “works,” whereas these laws preserve an identity. India can adopt the stance of providing publicity as astatutory right or have data protection laws similar to GDPR, in order to tackle AI conscious sound-alikes and protect performers’ rights for heirs of sounds like Kishore Kumar.

In mid2024, Drake released “Taylor Made Freestyle”, using AI-generated vocal impersonations of Tupac and Snoop Dogg, causing Tupac’s Estate to demand its removal.The 2023 proposed draft of NO FAKES Act in US, aimed to grant federal protection by making unauthorized digital replicas by AI subject to liability. Either model indicated the need for reforms at legislative and regulatory level to aim and safeguard vocal traits.

Recommendations and Conclusion

Section 57 of the Copyright Act does not provide any protection to performers from sound-alikes and look-alikes, and scraping cases like Amarnath Sehgal, or some limited protection of personality rights, like Amitabh Bachchan v. Rajat Negi are few and far between, and offer very limited solace.

The following three reforms are suggested in order to protect performers from unauthorized AI-generated imitation of their vocal identity. First, amend Section 57 and registration of a vocal style as a “work” to protect performers like Lata Mangeshkar against imitation by an AI being, may it be AI platforms like Audimee and Resemble AI. Second, codify personality rights by legislation akin to Midler v. Ford Motor Co., to create rights for performers and their heirs over the control of their identities. Lastly, laws that specify the uniqueness of deepfakes must be created, given the rise in their usage. India can draw from Denmark’s move of amending its copyright laws and carve out an exception for satire and irony. This system would mean accurately and statutorily defining what constitutes irony or satire which can bring in problems. Vague statutory definitions of satire and irony potentially can erode free expression, invite censorship and cause legal inconsistencies, demanding safeguards to democratic discourse.

Updates to the IT Act Section 66E could also deal with creating ‘imposter’ faces and voices when relying on AI. Cooperation among industry stakeholders to set ethical standards in the future may also be helpful, just as music labels worked with Meta to build safe detection of deepfakes. The ongoing case of Ani Media Pvt Ltd v. Open AI underscores the urgent need to protect performers’ rights through necessary legal protections. Without laws, performers would find it hard to rely merely on honor and reputation as a form of protection. Legislative reforms are therefore the need of the hour for India’s cultural icons and the soul of artistic expression.

Authored by: Vishesh Majumdar, IIIrd Year Law Stuent, Candidate Bennett University.

Edited by: Anushree Srivastava and Jhalak Bhatia