At the forefront of New Delhi’s India AI Impact Summit 2026, Amitabh Kant, former NITI Aayog CEO, issued a clarion call. India and other developing countries should urgently develop bespoke Large Language Models (LLMs) rooted in their own data. Why? Tech behemoths are siphoning Global South data to hone their AI, potentially locking benefits behind paywalls.
Kant cited compelling evidence: India’s data contribution surpasses the US by 33%. This influx is propelling LLM sophistication, but without countermeasures, it hands leverage to foreign entities crafting lucrative business models.
Equity demands action, Kant argued. AI needs to be accessible, ethical, and multilingual, fostering positive change across diverse populations. He flagged the peril of unchecked AI expansion amid soaring investments—it could entrench social divides.
Key challenges include reaching the world’s poorest, reshaping Global South realities, and advancing education, healthcare, and nutrition. AI’s magic lies in actualizing the improbable, from tailored teaching to optimized wellness.
Kant cautioned that ignoring the marginalized amplifies inequities. Instead, prioritize AI for outcome improvements in learning and beyond. This technology holds transformative power if steered inclusively.
Echoing through the summit, Kant’s blueprint for data sovereignty inspires a movement. Developing nations stand at a crossroads: innovate independently or perpetuate dependency. The stakes for global equity have never been higher.