Deepfakes And Criminal Liability: A Doctrinal, Comparative And Evidentiary Analysis
- IJLLR Journal
- 1 day ago
- 2 min read
Sampada Singh, MIT World Peace University, Pune
Ananya Agarwal, MIT World Peace University, Pune
ABSTRACT
The augmentation of artificial intelligence has validated the creation of deepfakes - synthetic media where artificial intelligence is used to realistically replace or alter a person's likeness, voice, or actions in images, audio, or video. The term is a portmanteau of "deep learning" (a type of AI) and "fake," and is commonly used to describe both the technology and the manipulated content. This have given rise to a unique and difficult question of criminal liability. This paper scrutinizes the doctrinal and substantive criminal law response to deepfake- enabled offences through four interlinked lines of inquiry. It first examines the doctrinal foundations of criminal liability; it examines how actus reus is fragmented across the chain of events, how mens rea must be evaluated across various actors with divergent knowledge and intent, and how causation, mode of participation, and corporate liability principles apply to AI-mediated harm. It also examines whether AI systems, lacking legal personhood, can meaningfully be a locus of criminal fault, or whether the doctrine of innocent agency better explains liability for AI-facilitated offences. The paper then undertakes a comparative survey of statutory frameworks across India, the United States, the European Union, the United Kingdom, evaluating each jurisdiction’s legislative approach to deepfake-related harm. It further analyses judicial responses, focusing on Indian judiciary’s extension to personal rights and privacy jurisprudence to synthetic media in the absence of a dedicated legislation, while noting a marked absence of criminal – as opposed to civil-injunction – precedent in India. Finally, the paper examines various challenges of prosecuting deepfake offences, including the fact that MLAT requests to foreign platforms/servers can take months to years, during which evidence may be deleted or altered, section 334-336 and section 356 of Bharatiya Nyaya Sanhita, 2023 exist as remedies but fail to adequately address the unique challenges of deepfakes — they were designed for physical/textual forgery and human-authored defamation, not AI-synthesised likeness. This paper concludes that while courts have demonstrated considerable doctrinal ingenuity in adapting existing law to synthetic media nonetheless persistent gaps in mens rea calibration and evidentiary standard continue to weaken the criminal law’s capacity to respond effectively to crimes facilitated by deepfake technology.
