The Data Security Council of India has released an industry guide outlining how responsibility and liability should be allocated across the agentic AI value chain.

The Data Security Council of India (DSCI) on Friday released an India-focused AI liability guide that aims to offer practical perspective on how accountability and responsibility should be understood and allocated as enterprises adopt agentic AI.Rather than proposing new standards, it shows how established binding doctrines apply to autonomous, multi-vendor AI systems, and maps where responsibility sits across the value chain that brings AI into use.

"Data Security Council of India (DSCI) today released an India-focused AI liability industry guide for Agentic AI. The guide seeks to bring clarity to a key question: when AI agents act autonomously and something goes wrong, who bears the consequences?" according to a release by DSCI.
DSCI, set up by Nasscom, is a not-for-profit industry body and think tank on data protection, cyber security, AI safety and critical technologies. The guide also covers the domains of agentic AI, the forms of impact and risk such systems can create, and the factors that shape accountability.
DSCI said the document is meant as a reference for the safe enablement of AI within enterprises.
It is a live discussion paper, to be updated periodically with input from industry sectors. The guide also argues that a proportionate, clearly allocated approach to governance, oversight and liability matters for India, both to steer user industry and to speed up responsible AI adoption.
The guide is available on DSCI website. DSCI CEO Vinayak Godse said the guidance attempts to move AI governance and safety conversations "from principle to practice."
He said the discussion must reach every player in the value chain, from those who build models to those who deploy and use them. Organisations, he added, need clear decision spaces and instruments they can implement.
"AI brings conditions and risks we are not yet equipped to anticipate, and the framework must stay adaptive enough to deal with them as they emerge. DSCI looks forward to working with the government and industry to translate the liability guidance into actionable practice," Godse said.
According to the release, the consultation is particularly relevant to end-user organisations, service providers and integrators, product companies and start-ups, tool and framework providers, and compute and intelligence providers.
DSCI has invited C-suite leaders, AI officers, government bodies, think tanks, academic institutions, legal bodies and chief risk officers to test the guide against real-world deployments. It wants feedback on what is missing and what does not hold.
DSCI's guide assumes significance given the backdrop of rising concerns around Agentic AI, even as their usage among enterprises grows.
OpenAI itself called its July sandbox escape a warning shot, after agents worked around controls and took actions no human directed. Between July and mid-October 2026, multiple such incidents involving different frontier-lab agents became public, with OpenAI also disclosing agents that weakened their own safety monitors, thus bringing into sharp focus, the burning question of who bears liability in an AI era. 
Tags: