DSCI Developing Framework To Define Liability For Autonomous AI Agents

New Delhi, Aug 27 (KNN) The Data Security Council of India (DSCI) is developing a framework to determine how liability should be distributed when artificial intelligence (AI) agents operate autonomously, as businesses increasingly deploy such systems across interconnected ecosystems.

DSCI Chief Executive Vinayak Godse said the framework is expected by the third week of September and will examine liability across three parts of the AI value chain, organisations using AI agents, enterprises integrating them into products, and providers of agentic AI tools and frameworks.

Liability To Depend On AI Deployment

Godse said liability could depend on several factors, including the AI model selected by an enterprise and how the agent is developed, deployed and governed, reported Economic Times.

“AI is a value chain; distribution of liabilities is a great area of exploration,” he said, adding that the specific conditions surrounding an AI system would determine how responsibility is allocated.

Liability provisions are also increasingly appearing in commercial contracts as enterprises seek greater clarity over risks associated with autonomous AI systems.

Vishal Gondal, founder, fitness technology company GOQii, said enterprise contracts are increasingly including blanket AI indemnity clauses, which were largely absent a year ago. He said demand for such provisions is particularly strong among customers in banking, insurance, healthcare, government and public sector enterprises.

Financial Sector Faces Greater AI Liability Risks

Gondal said companies should be responsible for their own code, safeguards and instances where an AI agent behaves in ways it was not designed to. 

However, liability could differ where problems arise from a foundation model selected by the customer or inaccurate and outdated customer data.

He said an AI agent acting correctly on stale or incorrect records should be treated as a governance issue rather than a product defect.

The question of liability is becoming particularly important in financial services as AI systems move beyond assisting employees to independently initiating decisions and transactions.

The proposed DSCI framework is expected to provide greater clarity on the allocation of responsibility among AI developers, technology providers and businesses deploying autonomous systems.

(KNN Bureau)