DSCI proposes liability framework for autonomous AI agent harms
A new industry paper sets rules for enterprise liability across five supply chain roles.
The Data Security Council of India (DSCI) has released a draft framework aimed at establishing legal liability protocols whenever autonomous artificial intelligence agents trigger damages [1].
Outlined in the discussion paper, ‘AI Liability: Industry Guidance for Agentic AI’, the guidance addresses the increasingly complex operational chains found within modern enterprise software deployments [1]. The framework segments systemic accountability across five distinct industry roles: end-user organizations, service providers, product companies, tool providers, and cloud and infrastructure providers [1].
DSCI's paper identifies 12 distinct categories of potential damage generated by agentic deployments, highlighting security threats, privacy breaches, and deepfakes among the key exposure areas [1]. As autonomous workflows expand across enterprise architectures, defining clear accountability boundaries between base infrastructure vendors and final end-user companies remains an unresolved challenge for corporate governance [1].
Questions and answers
What is the purpose of the DSCI framework?
It outlines guidance for assigning liability across five stakeholder groups when autonomous AI systems cause damage.
Which categories of harm does the guidance highlight?
The discussion paper identifies 12 areas of potential harm, including privacy breaches, deepfakes, and cybersecurity threats.
Companies & investors in this story
Sources
- 1.DSCI Floats Framework To Determine Liability For Harm Caused By AI Agents — Inc42, Nov 4, 2026
- 2.DSCI releases India-specific agentic AI liability guide — ET Datacenters
