DSCI Floats Liability Framework for Autonomous Agentic AI Harms

4 Nov 2026, 9:44 am · 22h ago · 1 min read · Inc42

The Data Security Council of India has released a discussion paper proposing an industry framework to assign liability for harm caused by autonomous AI agents. Titled ‘AI Liability: Industry Guidance for Agentic AI’, the paper suggests applying existing legal and contractual frameworks rather than creating new regulations. DSCI recommends distributing liability across the value chain—including end-user organisations, tool providers, and cloud infrastructure vendors—based on who controls key decisions and oversight. As agentic AI systems independently execute transactions, interact with APIs, and access enterprise data, the framework seeks to mitigate operational risks like breaches and unauthorized payments.