Peer reviewed analysis from world leading experts
Digital Futures Lab
Global AI governance has shifted from frontier risk mitigation towards development outcomes, with India’s 2026 AI Safety Summit centring the agenda around the Global South. While India has positioned itself as a responsible AI leader on the world stage, championing safety and development, its domestic AI governance framework relies on voluntary industry measures, optional incident reporting and institutions that exist in name only. While India’s contribution to shifting global AI debates towards Global South needs is valuable, it requires New Delhi to establish accountability structures to back it up at home.
When the United Kingdom hosted the first global AI Safety Summit in November 2023, mitigating frontier artificial intelligence (AI) risks dominated the agenda. Three summits later, the emphasis has shifted from what AI might do to humanity, to what AI could do for it. India’s 2026 AI Impact Summit, held in New Delhi in February 2026, focused on how AI can support development outcomes.
Safety did not disappear from the agenda — the New Delhi Declaration expressly recognises secure, trustworthy and robust AI as foundational to maximising economic and societal benefits. But this reframing marked a deliberate effort to centre the priorities of developing countries in global AI governance debates.
India has consistently sought to stake its claim as a global AI leader which takes responsible AI and risk mitigation seriously. It has aligned itself with a series of international commitments emphasising risk mitigation, cooperation and accountability — including the 2023 Bletchley Declaration, the 2024 Seoul Ministerial Statement, the 2024 UN resolution on AI risks and the 2025 BRICS Leaders’ Statement on AI governance.
The New Delhi summit produced substantive safety outcomes. The Safe & Trusted AI Working Group unveiled the Trusted AI Commons framework to pool governance and technical safety resources, context-specific datasets and benchmarking and evaluation frameworks across jurisdictions lacking such infrastructure. The Global South Network for Trustworthy AI was also launched, a civil society-led platform to assess AI’s real-world impacts and build locally grounded oversight mechanisms.
But for all its international positioning, India’s domestic AI governance is conspicuously short on enforceable action. India’s 2025 AI Governance Guidelines signal a clear preference for light-touch regulation, placing primary responsibility for risk mitigation on voluntary industry-led measures, while encouraging sectoral regulators to introduce binding rules as sector-specific risks emerge. Regulatory caution can help avoid locking in premature or poorly designed rules in a fast-evolving technological landscape, but the guidelines recommend making incident reporting optional, weakening the foundations on which any future accountability framework would depend.
India’s AI guidelines are more prescriptive at the sectoral level. The Reserve Bank of India has outlined expectations around product approval processes, consumer protection, AI-specific audits, cybersecurity and disclosure to consumers when they interact with AI systems. The National Health Authority also calls for risk-based classification of AI systems, along with validation checks and ongoing monitoring.
Yet none of these proposals, whether at the national or sectoral level, are backed by enforceable obligations. While some institutional movement has occurred, like the April 2026 constitution of the AI Governance and Economic Group and the Technology and Expert Committee, these have yet to demonstrate any meaningful function. And the IndiaAI Safety Institute, established as a domestic counterpart to global safety efforts in 2025, remains an aspiration rather than an operational body with resources, a mandate and independence.
The result is a persistent gap between principle and practice. Internationally, India champions trustworthy AI and positions itself as a leader in shaping global norms, while its domestic AI framework defers on key elements of accountability.
This gap becomes clearer in global comparison. Like India, Brazil is a large federal democracy grappling with inequality, institutional constraints and the challenge of balancing innovation with accountability. Though Brazil also debated whether comprehensive AI regulation was premature and relied initially on existing legal frameworks, it has since moved to advance a rights-based AI bill which introduces risk classifications, transparency obligations and penalties for non-compliance. While implementation and capacity challenges remain, this reflects a willingness to move from principles to enforceable obligations, with the understanding that waiting for perfect clarity risks entrenching regulatory inertia.
India, by contrast, continues to hedge. Its national framework defers to industry self-regulation even as sectoral regulators signal the need for stronger governmental oversight. New Delhi builds institutions in name but not function, endorsing global norms but stopping short of embedding them in enforceable domestic structures.
Reasons for this caution include concerns about stifling innovation, regulatory capacity constraints, federal complexity, the difficulty of governing a rapidly evolving technology across diverse sectors and the pressure to remain competitive in the global race to AI dominance. These considerations are not trivial. But they fall short of explaining the reluctance to establish even baseline requirements — mandatory incident reporting, enforceable transparency obligations when citizens interact with AI and a statutory footing for coordinating institutions with genuine independence.
India’s contribution to shifting the global conversation on AI towards development and inclusion matters. But helping to reframe the question is not the same as answering it. The Global South needs a leader that can show how AI delivers for people while remaining accountable to them. On the first, India has much to offer. On the second, it has yet to begin.
Shefali Malhotra is Senior Research Manager, AI & Tech Policy, Digital Futures Lab.
Your email address will not be published. Required fields are marked *
Sign up to receive free daily think pieces from leading analysts or our weekly digest, that includes our editorial and a collection of recent articles in brief.
Article printed from East Asia Forum (https://www.eastasiaforum.org)
Copyright ©2026 East Asia Forum. All rights reserved.
Did you know you can receive East Asia Forum analysis directly to your inbox?
Join our mailing list to never miss an update from top Asia analysts.
Subscribe
source
Leave a Reply