Borrowed Faces, Fabricated Voices and Legal Challenges of Deepfakes

Author

Nitish Kumar Nigam, Ansh Singhal, Sanidhya S. S. Chauhan

Abstract

 With all the attention and excitement over Generative Adversarial Networks (GANs), diffusion models and the like, deepfakes have moved on from a technological curio to something of major legal social and ethical concern across the globe. The article first surveys the international legal and regulatory response to those issues before tracing, on Indian law, two concurrent processes: slow-burning legislative and executive regulation at the national level culminating in 2026 Amendment to the Intermediary Rules; and a faster pace of development of judicial pronouncements over IP/personality-fears with High Courts extending copyright. Trade mark and privacy principles, by way of illustrations by case studies of which in broad strokes there are two, the Delhi high courts (Grok etc. and Chaitanya Rohilla) and Bombay high court (Rashmika) and in part Delhi as well the Bollywood 'star', including as a particular case study an analysis of the case of Getty Images v Stability AI, on training data and IPR for AI. The article then explains two paths of development of Indian law with regard to deep-fakes; it also makes a prediction that neither the 2026 Rules, albeit a move from warnings to statutory rule, would be sufficient to deal with the issue. 

Some such cases could be use of AI to impersonate celebrities (e.g. Anil Kapoor), the recent 'Rashmika Mandanna' issue and subsequent regulatory action. This could be contrasted with Chaitnaya Rohilla PIL and more generally with cross border issues such as the recent expanding 'Grok/xAI' litigation over various jurisdictions. Although the 2026 Rules are important step from soft advisories to hard law, they still cannot handle the depth of the issue and we are missing a number of essential aspects like a persona right and commensurate criminal offences; as well as a serious and credible mechanism for compensating victims.

Keywords

Borrowed Faces, Fabricated Voices and Legal Challenges of Deepfakes

Full Text:

Download Paper PDF

References

Three years separate the Rashmika Mandanna deepfake from the Grok controversy, and in that time, India has moved from a purely reactive advisory regime to a statutory, if still incomplete, framework under the 2026 Amendment Rules, propelled less by proactive policymaking than by sustained judicial pressure in litigation such as Chaitanya Rohilla v Union of India and a rapidly expanding line of personality-rights injunctions typified by Anil Kapoor v Simply Life India. The comparative picture shows India is not alone in improvising: the EU has chosen deployer-side transparency, the US is litigating its way toward product-liability standards through cases like the Grok lawsuits, the UK is testing the limits of copyright doctrine in Getty v Stability AI, and China alone has attempted a comprehensive, consent-first administrative scheme since as early as January 2023. What unites every jurisdiction, including India, existing legal categories, forgery, copyright, trade mark, tort, intermediary safe harbour, were not built for synthetic media, and each is being stretched, sometimes awkwardly, to cover harms it was never designed to address. Whether India’s 2026 Rules mark the beginning of a coherent statutory response, or merely the latest stopgap before the next viral controversy forces the next round of reform, the gaps are identified a free-standing personality right, calibrated criminal offences, cross-border enforcement capacity, and a genuine victim-redress mechanism, are addressed before, rather than after, the next crisis.

Share your valuable work from Social Media Buttons