IndiaAI Mission: Panel Calls For Setting Up Technical Secretariat To Govern AI

An advisory panel, formed under the IndiaAI Mission, has suggested sweeping reforms to the regulatory landscape to govern artificial intelligence (AI) systems in India and ensure the trustworthiness and accountability of such platforms. 

A sub-committee on ‘AI Governance and Guidelines Development’ formed under the panel’s guidance has recommended establishing an inter-ministerial AI coordination committee to bring together various authorities and institutions which deal with AI governance at the national level. 

As per a draft report floated by the panel, the coordination committee should be entrusted with applying and strengthening existing laws to minimise risk of harm due to use of AI, providing legal clarity and creating a policy environment that enables the use of AI for beneficial use-cases without impeding innovation. 

Calling for a “whole-of-government” approach to AI governance, the recommendations also call for creating sector-specific datasets to enable the creation and evaluation of “fair” AI models. 

The draft report says that the proposed coordination committee should be headed by the principal scientific advisor, adding that its members should include representatives from the Reserve Bank of India (RBI), electronics and IT ministry (MeitY), NITI Aayog, other relevant departments, and industry stakeholders. 

It is pertinent to note that the union government selected eight strategic projects, in October last year, under the INR 10,371 Cr IndiaAI Mission to promote responsible development, deployment and adoption of AI in the country.

Subsequently, the advisory panel, under the chairmanship of the principal scientific advisor, was set up to develop an India-specific AI regulatory framework. The panel then tasked a sub-committee to provide recommendations and examine key issues related to AI governance in the country.

Now, the draft rules have been floated for public consultations. The last date for providing comments or feedback is January 27.

The advisory panel has also called for establishing a sub-group to work with MeitY to suggest specific measures that can be considered under proposed legislations, such as the Digital India Act (DIA), to harmonise the legal AI framework, technical capacity and adjudicatory mechanisms for effective grievance redressal and ease of doing business.

Making Way For MeitY Technical Secretariat

The advisory committee has recommended setting up a “technical secretariat” within MeitY to aid in AI governance in the country. 

It has pitched for the following functions for the technical secretariat:

• Pooling together multi-disciplinary expertise (tech, law, policy, economics, among others) to strengthen capacity across departments and regulators
• Creating a map of the stakeholders and actors involved in India’s AI ecosystem and conducting regular horizon-scanning exercises of the AI field
• Assessing risks to consumers and society across applications and domains 
• Facilitating the development of metrics and common frameworks (such as data provenance, system cards, security, among others)
• Engaging with industry to co-examine novel solutions to ensure responsible use of AI and enable development of appropriate guardrails
• Identifying gaps which may not be adequately addressable through delegated legislation 

The recommendations also note that the proposed technical secretariat should establish an AI incident database and private enterprises should be encouraged to report such cases. 

“The AI incident database should not be started as an enforcement tool. Its objective should not be to penalise people who report AI incidents. Instead, the objective should be to encourage reporting and the learnings should flow back into the ecosystem. Given this, the suitability of CERT-IN taking on the mandate of maintaining an AI incident repository, under the guidance of the Technical Secretariat, may be examined,” as per the draft report .

The advisory panel has also said that the technical secretariat should examine the suitability of technological measures and solutions to address AI related risks to complement the legal framework.