Deepfakes, Criminal Responsibility and the Limits of India’s Cybercrime Framework: Rethinking Liability in the Age of Generative AI

Abstrakt

Generative artificial intelligence has transformed the production of synthetic audio-visual content, making it increasingly difficult to distinguish authentic recordings from artificially generated or manipulated representations. Deepfakes present a distinctive challenge for criminal law because the technology can facilitate fraud, impersonation, extortion, non-consensual sexual imagery, reputational harm, and other forms of digital wrongdoing while simultaneously complicating the attribution of criminal conduct to a particular human actor. This article examines whether India's existing criminal-law framework is capable of assigning responsibility for unlawful conduct involving deepfakes, with particular attention to the Bharatiya Nyaya Sanhita 2023, the Information Technology Act 2000, the Bharatiya Sakshya Adhiniyam 2023 and the evolving intermediary framework governing synthetically generated information. The article adopts a doctrinal and comparative methodology, supplemented by an analysis of the emerging relationship between artificial intelligence, mens rea, attribution and criminal responsibility. It argues that the principal difficulty is not simply the absence of a dedicated deepfake offence, but the fragmentation of existing legal responses and the uncertainty surrounding the attribution of culpability where human conduct is mediated by generative AI. The article develops a differentiated framework based on human agency, intention, knowledge, control and resulting harm, distinguishing the criminal responsibility of creators, users, disseminators and intermediaries. It further examines comparative approaches in the European Union, United Kingdom and United States to identify principles that may inform Indian law without mechanically transplanting foreign regulatory models. The article concludes that India's response should preserve the foundational principle of individual criminal responsibility while developing technologically informed rules for attribution, proof and differentiated liability. Such an approach can address serious deepfake-enabled harms without converting every act involving synthetic media into a criminal offence.