By Sunil Mehta and Richa Roy

In the six weeks that preceded the BRICS Summit in New Delhi on September 12 and 13, both Demis Hassabis and Bill Gates underscored the need for an AI governance institution, tracking risks across labour, finance, elections, health and beyond. Both emphasised international coordination; yet did not envisage fast-growing emerging markets represented in such institutions.

Advertisement

AI’s risks, however, transcend sovereign boundaries and sectoral silos. The chair of the Financial Stability Board (FSB), Andrew Bailey, warned that leverage, stretched valuations and AI optimism are interacting in ways that could amplify a market correction across borders, implying a bubble, a global financial and economic risk. Crucially, emerging markets are massive users of AI; their resources, data and labour power its models, and they are most vulnerable to its technological, sovereignty, financial and climate risks. The BRICS Summit presents an opportunity to address them.

When India held the 2023 G20 presidency, the authors argued for an International Institute for Regulatory Development (IIRD): Under the aegis of G20 and FSB, to set standards for emergent technologies, undertake global regulatory research, and “tropicalise” rules for fast-growing emerging economies rather than transplant them from the North Atlantic regulatory construct. Three years on, AI sharpens the argument for IIRD.

What New Delhi BRICS Summit can build

The IIRD should be a global apex body for AI, finance and climate risk, with three core functions.

Advertisement

First, research, regulatory capacity building for early warning. A permanent secretariat producing shared assessments of how AI and other global risks transmit through financial systems and climate exposures, with data drawn from emerging as also advanced markets. Existing institutions are inadequate – the FSB is a coordinating umbrella, not a research institute, and the two should complement, not duplicate each other.

Second, an exchange of practice. A living repository of what has worked. The precedents exist. AGRORA, a project with 40 private participants such as SWIFT as well as central banks of G7 plus countries as well as Korea, Mexico and others and Project Dunbar, in which the central banks of Singapore, Australia, Malaysia and South Africa built a shared platform for digital currencies, shows what global regulators can build together rather than wait. The results of AI pilots in UPI could, for instance, be a valuable part of this exchange.

Third, proportionate standards. Weighted governance that reflects where growth and users are, and a formal channel into the FSB, the Basel Committee, and climate and frontier-AI bodies. Proportionality matters as much as representation: Standards to be calibrated to contexts.

India’s BRICS Chairship Summit, with AI governance on the agenda, is the natural incubator. BRICS now spans eleven members and ten partners.

India’s role

Given the geopolitical complexity, including concerns around specific nations, design choices are critical. The IIRD should be seeded at Delhi, though not owned by BRICS: Chartered as an open institution, housed in India, funded by a mix of public, private and multilateral funds, with the New Development Bank as one contributor among several, and with membership open to G20 (UK’s forthcoming Presidency of G20 suitably places it to raise evolving global risks, including AI governance; it was also the first host of the AI Summit at Bletchley), the OECD, multilateral development banks and the FSB.

The IIRD would not threaten the turf of existing institutions, but complement them by producing shared research and early warning on how AI risk transmits through finance and climate, with data from emerging and advanced markets — a gap the FSB itself admits; facilitate a living exchange of practice; feed proportionate, mutually recognised standards into the FSB, Basel; and be a precursor to a global frontier-AI body.

The Delhi declaration should mandate a working group, chaired by India and open to non-members, to bring an IIRD charter to the 2027 summit, and to present it in parallel to the G20 finance track: Critically, to institutionalise a communique.

India’s presidencies of global groupings have been praised for being inclusive (adding the African Union as a permanent member of the G20). As with any international summit, there is the risk of (perhaps unwarranted) critique of the spectacle of the event surpassing the technical negotiations and sustained institutional engagement which are the real backbone of international cooperation. The BRICS IIRD working group offers an opportunity to both retain the inclusivity and preclude the critique.

A third draft

Bretton Woods institutions originated in America and Europe, designed for a world that no longer exists. AI’s governance is currently being designed by a handful of companies and governments for a world that does not yet exist. Both exclude countries whose citizens will be AI’s largest, most vulnerable stakeholders. India’s common law heritage: Technological prowess and its Global South convening power (without ceding autonomy) allow for designing a different architecture. Its BRICS Chairship is the moment to offer a third draft, with the Global South as co-creators of the new rules. India should not let this opportunity pass.

Mehta chairs the Asia Society India Centre and is the former Chairman of PNB, YES Bank, IndusInd Bank and Country Head of AIG India. Roy is a Chevening Gurukul Fellow at the University of Oxford and partner at Cyril Amarchand Mangaldas. Views expressed are personal