Indic Pacific Releases "Allocating Liability for Harms Caused by Agentic AI Systems" by Nikitha Kotteswaran

Indic Pacific Legal Research has released a new monograph, Allocating Liability for Harms Caused by Agentic AI Systems: The Limits of Tort, Regulatory and Institutional Design (IPLR-IG-017), authored by Nikitha Kotteswaran. The monograph carries a foreword by Sunil Jose, Founder of SUNS Legal and Professor of Practice at Alliance University, and an introduction by Abhivardhan, President of the Indian Society of

Artificial Intelligence and Law.

Read and purchase the monograph here: https://indopacific.app/product/allocating-liability-for-harms-caused-by-agentic-ai-systems/

AI systems no longer only recommend. They act. When an agent commits a user to a non-refundable itinerary, pushes a breaking change to a production environment, or operates a home appliance that causes damage, existing liability rules do not give a clean answer on who pays. This monograph takes that question directly.

It defines the class of systems it covers with care: LLM-based agents that pursue goals across multiple steps, call external tools and APIs, and take consequential actions without a human approving each step. It then tests negligence, product liability and vicarious liability against these systems, and separates genuine doctrinal gaps from familiar rules applied to unfamiliar facts.

The monograph compares the regulatory responses of the European Union, the United Kingdom, the United States, Singapore, South Korea, India and China, and evaluates three candidate liability frameworks: strict developer liability, fault-based deployer liability, and a risk-tiered hybrid. It commits to a position. Under the recommended hybrid model, liability is calibrated to a system’s autonomy and the reversibility of its actions, backed by layer-specific statutory recommendations for developers and deployers, including mandatory insurance, incident reporting within 72 hours, and a qualified safe harbour for open-source developers.

Three case studies, a consumer transactional agent, an enterprise coding agent and a home robotics agent, show how each framework allocates liability in practice. A dedicated discussion covers what the monograph calls the insurance precondition: no strict liability regime is financially sustainable without a functioning insurance market behind it.

For India, the monograph reads a governance landscape fragmented across more than 30 instruments and proposes a phased response: what can be done immediately, what requires further institutional development, and what would eventually need a dedicated legislative framework.

As Abhivardhan writes in his introduction, “A monograph is an invitation to argue. This one gives practitioners, regulators, and scholars something solid to argue with.”

About the author

Nikitha Kotteswaran is a legal researcher working at the intersection of technology, finance and public policy. She is pursuing an LL.M. in Investments and Securities Law at Maharashtra National Law University Mumbai and the National Institute of Securities Markets, and serves as a Research Fellow in Fintech at the Centre for Development of Intellectual Property and Research (CDIPR).

Publication details

First Edition, October 20, 2026. ISBN 978-81-686447-4-8. Price: INR 200. The monograph is available on IndoPacific.App, and is published under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International licence.

Suggested citation: Nikitha Kotteswaran, Allocating Liability for Harms Caused by Agentic AI Systems (Indic Pacific Legal Research, October 20, 2026).

For correspondence, write to research@indicpacific.com.