DSCI releases AI liability guidance to address accountability risks in agentic AI deployments
The Data Security Council of India (DSCI) has released an industry guidance document on AI liability, outlining how accountability and responsibility could be assigned when autonomous AI agents cause errors, operational disruptions or other adverse outcomes.
Released on October 30, 2026, the India-focused guidance examines how existing legal provisions and established liability principles apply to agentic AI systems, particularly those operating across multiple vendors and platforms. It aims to help enterprises understand the responsibilities of different stakeholders as they integrate autonomous AI systems into business operations.
Addressing accountability in autonomous AI systems
Unlike conventional AI applications, agentic AI systems can independently execute tasks, make decisions and interact with other systems with varying levels of human intervention. This raises questions about who should be held accountable when an AI agent takes an incorrect action or causes harm, particularly when multiple parties are involved in building, deploying and maintaining the system.
The guidance does not propose new standards or regulatory requirements. Instead, it maps existing binding legal doctrines to the emerging agentic AI ecosystem and examines how responsibility may be distributed across the AI value chain.
Guidance maps responsibilities across the AI value chain
The document covers different domains of agentic AI, the potential risks and impacts associated with their deployment, and the factors that could influence liability. It also highlights the need for clearly defined governance structures, human oversight and accountability mechanisms as enterprises move towards greater AI autonomy.
The document is intended as a reference for organisations adopting AI and is positioned as a live discussion paper. DSCI said it would update the guidance periodically based on feedback from different industry sectors.
DSCI calls for industry participation
Vinayak Godse, CEO of DSCI, said the guidance seeks to move AI governance and safety discussions from broad principles to practical implementation. He highlighted the need for accountability across the AI value chain, including model developers, deployers and end users, along with clearly defined decision-making responsibilities and mechanisms that organisations can implement.
The consultation is aimed at end-user organisations, service providers, system integrators, product companies, start-ups, tool and framework providers, and compute and intelligence providers.
DSCI has also invited industry leaders, including C-suite executives, AI officers, chief risk officers, government bodies, academic institutions, research organisations and legal professionals, to assess the guidance against real-world AI deployments and share feedback on gaps and implementation challenges.
As enterprises increasingly experiment with autonomous AI agents, defining accountability across the technology ecosystem is becoming an important part of managing deployment risks. The guidance seeks to provide an industry-level starting point for addressing these questions in the Indian context.