Author: Vanshika Shukla

ABSTRACT

Modern tort doctrines such as negligence, strict liability, and vicarious liability provide foundational tools for addressing harms caused by artificial intelligence (AI), yet they struggle to accommodate the unique features of autonomous, opaque, and continuously evolving systems. This paper employs a combined doctrinal, comparative, and normative methodology to argue that a hybrid liability architecture is the most appropriate pathway for India. Specifically, this paper advocates for: (i) strict liability for physical, safety-critical harms caused by autonomous AI systems; (ii) a calibrated fault-based regime for informational and reputational harms; and (iii) statutory algorithmic duties supported by procedural and institutional reforms. Drawing on Indian constitutional principles, landmark domestic and international jurisprudence, and the EU Artificial Intelligence Act (EU AI Act), the paper identifies structural doctrinal failures in existing law, critically evaluates the applicability of European regulatory models to the Indian legal context, and proposes concrete legislative provisions tailored to India’s socio-legal environment.