Part 4/13:
The panel explored the distinction between inventing foundational AI models versus applying existing technology to solve local problems. Nandan Nilakani pointed out that India doesn't need to build foundational models from scratch; instead, leveraging open-source models and adapting them contextually can be highly effective.
For example, a startup in Maharashtra developed a voice assistant that communicates in Marathi with farmers about weather and irrigation advice, resulting in a 65% increase in crop yield. Such use cases demonstrate how India is building impactful AI solutions across sectors like agriculture, healthcare, and education by focusing on vernacular languages and offline usability.