On the anvil: Broader AI risk framework for cos launching new services

The proposed framework is expected to set guardrails for AI tools, create a regulatory sandbox and define accountability norms for companies operating in India

The government is likely to introduce a broad artificial intelligence (AI) risk framework for companies launching new AI-enabled services in the country. It is likely to be part of a new artificial intelligence (AI) law in the works and could be presented to Parliament for approval within this year, sources told Business Standard.

 

The framework would define the broad contours, including dos and don'ts for AI tools and services, and establish a regulatory sandbox within which such products should be developed and operated, according to a senior government official.

 

“The AI risk framework will also establish the parameters that frontier AI development companies must follow when dealing with Indian users. The use cases in the country often differ very sharply from those in other parts of the world,” the official said.

 

An AI risk framework is also necessary to provide the Centre legal authority to establish accountability when AI tools and services are found to be in violation of prescribed norms, another official said. While the framework will be largely prescriptive, it will give the Centre certain emergency powers to issue directions for stopping a particular AI tool or service.

 

The work on the AI risk framework, along with the proposed AI law, is being carried out by the newly constituted Artificial Intelligence Governance and Economic Group (AIGEG).

 

This inter-ministerial body coordinates AI policy across key ministries and their respective departments, as well as sectoral regulators. It is also responsible for driving AI innovation that balances safety, accountability and the ethical deployment of AI tools and services.

 

Formed in April this year, the AIGEG is also expected to sharpen India's broader AI strategy by funnelling the bulk of funding into a limited number of use cases capable of delivering measurable outcomes within the next 12 to 18 months. Priority sectors are likely to include healthcare, agriculture and education.

 

The need for an apex AI governance body in India comes amid growing concern among experts over the rapid advancement of large language models (LLMs) and tools such as Claude Mythos, which have demonstrated the ability to penetrate legacy software systems and existing digital defences.

 

“While respective government departments are continuing the work of shoring up cybersecurity after the arrival of Mythos, there also needs to be a unified response mechanism,” said one of the officials quoted above.

 

India currently lacks a unified law on AI. The Information Technology Act, 2000, and the Information Technology Rules, 2022, govern most AI-led tools and services in the country.