India's Consumer Law Faces Gaps for AI-Related Harms

As AI systems become ubiquitous in consumer applications, the question of liability for AI-induced harms remains largely unresolved. This working paper analyzes whether India's Consumer Protection Act, 2019, is adequate for addressing damages caused by defective AI products and services and for proportionally assigning liability across the intricate AI value chain. The paper notes that the Act's expansive definitions of product liability, harm, and deficiency appear broad enough to cover various AI-related incidents, including physical injury, psychological distress, biased outputs, and loss of control. However, it identifies critical shortcomings. Proving a direct causal link between an AI defect and consumer harm is technically challenging, as AI failures often stem from design choices rather than discrete manufacturing flaws. Furthermore, the Act's traditional framework assumes distinct roles for manufacturers, sellers, and service providers, which does not neatly map onto the overlapping responsibilities of data providers, model developers, deployers, and users within the AI ecosystem. Consequently, existing liability frameworks lack the mechanisms to effectively address complex, multi-stakeholder AI harms in a proportionate manner.