Artificial Intelligence (AI) is transforming the functioning of legal systems by assisting in legal research, document review, predictive analysis, administrative decision-making, and other professional tasks. While these technologies improve efficiency, reduce costs, and support faster decision-making, they also raise complex questions of legal accountability when an AI system causes harm. This paper examines the central issue of liability for AI-related harm under Indian law, particularly whether responsibility should rest with developers, manufacturers, deployers, users, or, in limited theoretical discussions, the AI system itself. It analyses the application and limitations of existing legal principles, including criminal liability, civil liability, negligence, product liability, absolute liability, and vicarious liability. The paper also evaluates whether artificial intelligence can be recognised as a legal person under Indian law and explains why the absence of intent, conscience, and moral judgment makes such recognition difficult within the current legal framework. The study adopts a doctrinal and comparative approach by examining Indian statutes, case law, scholarly writings, and international developments, particularly the European Union Artificial Intelligence Act, 2024. It argues that existing Indian laws, including the Information Technology Act, 2000 and the Consumer Protection Act, 2019, were primarily designed for human conduct and conventional products, and therefore do not adequately address autonomous or algorithmic decision-making. The paper recommends that India enact a dedicated AI liability framework, amend consumer protection law to cover AI-enabled goods and services, introduce certification and oversight mechanisms for high-risk AI systems, and develop compensation models for serious AI-caused harm. Such reforms are necessary to protect victims, ensure accountability, and promote responsible innovation.