Authored by Dr. Pavan Duggal, Advocate, Supreme Court of India, Architect, Global AI Accountability
The International AI Accountability Forum convenes at New Delhi on 14 May 2026 to advance the New Delhi foundational AI Principles - the first Global South-authored framework for international AI accountability. The asymmetries of the prior century must not be reproduced in algorithmic form.
For more than three decades, every major architecture of international technology law has been written substantially in capitals where the Global South was, at best, consulted, and more often simply notified. That asymmetry must end with the law of artificial intelligence. It must end now, in 2026. And it must end on terms set in part by the jurisdictions that constitute the demographic majority of the planet.
What I have, in recent work, termed algorithmic colonialism is the export of governance-free AI into developing markets: the deployment of consequential algorithmic systems in jurisdictions whose regulatory institutions are structurally less able to interrogate them, whose constitutional frameworks were not contemplated in their design, and whose citizens are reduced from rights-bearing persons to extractable data subjects.
Algorithmic colonialism is not a rhetorical flourish. It is the operating dynamic of the current AI deployment cycle. A model trained in one jurisdiction, hosted in a second, sold by a multinational enterprise headquartered in a third, and consumed by a citizen of a fourth generates harms whose remedy, if any, is theoretical for the citizen.
The inadequacy of the present international architecture
This is the fundamental inadequacy of the present international AI architecture. The European Union’s Artificial Intelligence Act protects European persons. The state-level mosaic in the United States protects American persons. China’s algorithmic regulations operate within China’s distinctive constitutional order.
The OECD Principles, the UNESCO Recommendation on the Ethics of Artificial Intelligence, the Bletchley Declaration of November 2023, and the Seoul Declaration of May 2024, each valuable in its own register, establish norms without mechanisms of enforcement against private deployers operating across borders.
The Council of Europe Framework Convention on Artificial Intelligence, opened for signature in September 2024, is an important inter-state instrument. It does not, in its operative architecture, address the citizen of Lagos, of Manila, of Dhaka, or of Delhi who has been wrongfully denied credit, wrongfully arrested by an algorithmic match, or wrongfully classified by a clinical decision-support system imported from elsewhere.
The New Delhi Foundational Principles on AI Accountability, advanced for adoption at the International AI Accountability Forum convening at New Delhi on 14 May 2026, are offered as a corrective. They are not a regional instrument. They are a Global South-authored framework designed for international portability, anchored in the Duggal Doctrine of Ten Common Legal Principles for AI Regulation, and intended to supply, for the first time, a coherent normative architecture in which the developmental jurisdictions of the world are co-architects of the law rather than passive recipients of it.
Sovereign AI Accountability
At the centre of the New Delhi Foundational Principles is the doctrine of Sovereign AI Accountability: the proposition, advanced first at the Global Summit on AI, Emerging Tech Law and Governance (GSAIET 2025) and now operationalised, that every national jurisdiction retains the sovereign power, and bears the corresponding sovereign responsibility, to regulate AI systems deployed within its borders, regardless of the domicile of the developer or the deployer.
Sovereign AI Accountability is the constitutional anchor of the framework. It rejects the implicit premise of the current global order that AI systems travel under the regulatory regime of the jurisdiction in which their developers reside. That premise is incompatible with sovereign equality. It is incompatible with the constitutional traditions of every state in the Global South. It must be displaced.
India is positioned, distinctively, to anchor this displacement. The Constitution of India, in its commitments to equality before the law, to the right to life with dignity, to social and economic justice, and to fraternity, supplies a substantive ethical basis for an accountability framework. India’s experience as the world’s largest democracy, its most populous nation, a major AI deployer and a major AI-affected jurisdiction, supplies practical authority.
India’s convening capacity, demonstrated through the India AI Impact Summit at Bharat Mandapam in February 2026, where the New Delhi Declaration on Responsible Artificial Intelligence was endorsed by eighty-six countries and two international organisations, supplies institutional weight.
The Information Technology Act of 2000, to which I had the privilege of contributing foundational inputs, and the Digital Personal Data Protection Act of 2023 supply legislative foundations. None is sufficient on its own. Together, they constitute a platform from which a Global South-led international framework can be authored.
A framework offered, not imposed
The framework advanced through the New Delhi Foundational Principles is not directed against any jurisdiction. It is offered for adoption across legal traditions, levels of development, and constitutional cultures. Its principles are common, in the sense that they are designed to be portable across legal systems. Its architecture is enforceable, in the sense that it translates aspiration into doctrine. Its commitments are equitable, in the sense that they reframe the Global South from object of regulation to co-author of regulation.
The civilisational stakes are unambiguous. The law of artificial intelligence is being written now, in this decade, in instruments that will define the relationship between human dignity and machine cognition for generations. If that law is written without the meaningful participation of the jurisdictions that host the demographic majority of humanity, the asymmetries of the prior century will be reproduced in algorithmic form, and the international order will pay the cost in legitimacy, in stability, and in justice.
The International AI Accountability Forum 2026 is convened to ensure that this does not happen. The Global South must not be a passive recipient of AI governance norms designed elsewhere. It must be a co-architect of the law that governs the algorithms that increasingly govern its peoples. The work begins at New Delhi on the fourteenth of May. Let it begin well.
𝐒𝐭𝐚𝐲 𝐢𝐧𝐟𝐨𝐫𝐦𝐞𝐝 𝐰𝐢𝐭𝐡 𝐨𝐮𝐫 𝐥𝐚𝐭𝐞𝐬𝐭 𝐮𝐩𝐝𝐚𝐭𝐞𝐬 𝐛𝐲 𝐣𝐨𝐢𝐧𝐢𝐧𝐠 𝐭𝐡𝐞 WhatsApp Channel now! 👈📲
𝑭𝒐𝒍𝒍𝒐𝒘 𝑶𝒖𝒓 𝑺𝒐𝒄𝒊𝒂𝒍 𝑴𝒆𝒅𝒊𝒂 𝑷𝒂𝒈𝒆𝐬 👉 Facebook, LinkedIn, Twitter, Instagram
