AI that serves all: Reviving the spirit of Nalanda for a planetary commons

We both remember the first time we walked through the ruins of Nalanda together, in spirit if not in person. Fifteen centuries ago, that university in present-day Bihar was one of the world’s most remarkable centres of learning.

It drew students and scholars from across China, Korea, Central Asia and South-East Asia, who lived, argued and learned side by side. They studied logic, medicine, astronomy and the art of debate.

At the heart of Nalanda’s intellectual life was a Buddhist commitment to the union of wisdom and compassion: the belief that knowledge gains its highest meaning when it is directed towards reducing suffering and deepening understanding among all beings.

This ethos shaped Nalanda’s scholarly culture and continues to define its enduring legacy. Today, as we look at the implementation of India’s national AI Mission as well as the AI Hub for Sustainable Development in Africa, the lessons from Nalanda continue to hold true.

The challenges facing AI in 2026 are the very ones Nalanda answered in the 5th century: how to build knowledge systems that serve all of humanity, especially those historically left behind. Indeed, as the UNDP Human Development Report 2025 points out, people and their choices can and will shape the future of AI.

As Nandan Nilekani, co-founder and chairman of Infosys, recently wrote, "The hardest places to make artificial intelligence work are also the places where it matters most." A farmer in Vidarbha or Turkana needs an answer he can trust when the rains fail. A health worker in rural Bihar or northern Kenya needs a diagnosis she can act on today.

That trust does not come from more parameters; it comes from credible sources, cultural grounding, and the ability to question and correct the system – the institutional accountability Nilekani describes. This reflects the spirit of open debate institutionalized by Nalanda, recognizing that safe impact at scale requires deep, contextual collaboration across the foundational enablers of safety, models, multilingual and voice capabilities, compute and other horizontal unlocks focused on both people and the planet.

This is why India’s AI Mission is enabling access to GPUs at the lowest possible cost, why Bhashini – India's Ministry of Electronics and Information Technology language platform – already reaches millions of citizens in their own languages, and why voice-first applications are being built for the half-billion Indians still offline.

Co-designed by the G7 during Italy’s Presidency with Kenya, Egypt and other countries, the AI Hub for Sustainable Development is working with innovators to break silos to catalyse a historic opportunity to move from raw-material exporter in Africa to high-value AI-powered industrialized economies by bundling critical minerals and renewable energy with modular, right-sized compute, connectivity, micro data centres and data pipelines.

In both places, we are deliberately designing fairness and equitability of access, voice and ownership into the horizontal unlocks or the stack. We are surfacing new win-win frameworks rooted in collaboration and development impact for people.

This fundamental insight echoes across a diverse tapestry of global wisdom traditions. Like Nalanda, the Vatican’s Rome Call for AI Ethics affirms that technology must serve human dignity and the common good, never domination. Indigenous African philosophies of ubuntu remind us that "I am because we are." Confucian traditions stress harmony between human intention and technical power. All converge on one truth: exclusive knowledge systems become extractive; inclusive ones are profoundly generative.

In the age of AI, the global majority are surfacing the many horizontal enablers and starting to collaborate across a network of organizations to power it – whether it is investing in modular renewable powered data centres, where computing infrastructure is integrated with local payment systems to ensure wider access and affordability, or building strong community stewardship of AI's linguistic and cultural diversity that make services valuable to a farmer or a microentrepreneur and, in turn, shapes global benchmarks in AI to address bias and misinformation. For example, how AI-driven climate intelligence, which combines geospatial data and climate trends, can guide crop planning, water management and public investment – work that UNDP India supports through DiCRA, alongside the Government of India and NABARD.

The debates at Nalanda were fierce, and every scholar, regardless of origin, had to defend their ideas in public. We are trying to rebuild that spirit for AI. When a new model is deployed in a government school in Odisha or a clinic in Côte d’Ivoire, teachers and health workers can flag errors, suggest improvements, enabling model retraining in weeks, not years.

When a farmer in Maharashtra speaks to MahaVISTAAR or a pastoralist in Turkana speaks to a Tabiri-powered early-warning system, the answer is traceable back to an agricultural university or meteorological institute. This is Nalanda’s logic translated into code: transparency, accountability, and continuous collective input. Because impact happens only when a real problem is solved.

Yet too often, AI efforts splinter across teams, vendors and programmes, each solving a local problem with its own data, infrastructure and metrics. The result is islands of progress that fail to reach scale: every team rebuilds the stack, and momentum dies before anything composes at national or continental levels.

True AI value emerges only when people actually use it in everyday life, shifting from isolated pilots to widespread adoption.

Today, most efforts stall after initial deployment – not because of a technology gap, but because of a coordination and governance gap across the horizontal enablers.

The economic stakes are enormous. AI could add $15.7 trillion globally by 2030 and $1.5 trillion in Africa alone. AI today is what electricity and the internet were in their early days—a general-purpose technology with the power to reshape every industry, every workflow and every society.

Just like those earlier revolutions, AI opens up many possibilities. But possibilities alone don't create impact. Only implementation does.

The moral stakes are even higher. If we fail to transition to a fair AI, we deepen the divide between those who own the models and those who are merely modelled. But if we get it right, we create the first truly planetary knowledge commons since Nalanda itself.

In 2026 India will host the AI Impact Summit, the first major global AI convening led by the Global South. We want it to feel like Nalanda reborn: open gates, fierce debate, shared meals and a single shared purpose: to prove that artificial intelligence can be the most powerful equalizer humanity has ever built.

Because if AI can work in the classrooms of rural Bihar and the pastures of northern Kenya, it can work anywhere. And if it serves the farmer and the health worker first, it will serve everyone.