New Delhi, Feb. 22 -- For much of the world, the barriers to artificial intelligence are not philosophical but practical: the cost of computing power, the absence of reliable internet, the lack of trained workers. The New Delhi Declaration on Artificial Intelligence, adopted on Saturday by 88 countries and international organisations, puts those concerns at the centre of global AI governance for the first time. The declaration argues that "robust digital infrastructure and meaningful and affordable connectivity are prerequisites for deploying AI and unlocking its full potential," calls for "upgrading vocational and training ecosystems" and the "training of public officials" in AI, and frames affordable AI systems as a means to "accelerate local innovation" in developing economies. It endorses open-source AI - qualified with the phrase "where appropriate" - as a tool for scalability across regions that cannot build proprietary systems from scratch. The AI Impact Summit 2026, hosted by India, is the first in the four-summit history of global AI governance to be held in the Global South - and the shift in geography is reflected in the shift in emphasis. "The advent of AI marks an inflection point in the trajectory of technological evolution. The choices that we make today will shape the AI-enabled world that future generations will inherit," the declaration stated. To be sure, the declaration carries no legal force and does not obligate signatory countries to specific legislative or regulatory action. Previous summits in Bletchley Park, Seoul and Paris were shaped primarily by the preoccupations of technologically advanced Western nations: existential risk, safety standards, regulatory frameworks. New Delhi has not abandoned those concerns - a full pillar of the declaration is devoted to secure and trustworthy AI - but it has subordinated them to a prior question: who gets to participate in AI at all. The declaration was endorsed by 88 countries and international organisations, the broadest sign-on yet. It surpasses the 58 signatories at last year's Paris summit, where the United States and the United Kingdom declined to participate. Both have signed in New Delhi, alongside China, Russia and the European Union. "[The] entire world has endorsed PM [Narendra Modi's] Ji's human-centric vision of AI. The declaration is inspired by the principle of 'Sarvajana Hitaya, Sarvajana Sukhaya', for democratising AI resources for the global population," said Union minister for electronics and IT, Ashwini Vaishnaw. The summit announced seven voluntary, non-binding frameworks, including a Charter for the Democratic Diffusion of AI to promote "access to foundational AI resources" and "locally relevant innovation"; a Global AI Impact Commons to replicate successful AI use cases across regions; a Trusted AI Commons of "technical resources, tools, benchmarks and best practices"; and an International Network of AI for Science Institutions to pool research capabilities globally....