Nilekani highlighted the need for high-quality datasets, especially in Indian languages, to support AI growth in India