At the India AI Impact Summit 2026 held in New Delhi, Pratyush Kumar, co-founder of Sarvam AI, delivered a strong message about the future direction of artificial intelligence in India. He emphasized that if India aims to become a global AI leader, it must move beyond English-centric AI development and build systems that reflect the country’s vast linguistic and cultural diversity.
Kumar highlighted that most global AI models and benchmarks are primarily designed around English language datasets and Western contexts, which limits their relevance in a country like India, where hundreds of languages and dialects are actively spoken. He stressed that true AI inclusion means designing models that understand regional languages, local accents, rural realities, and culturally specific user behaviors. Without this shift, AI risks remaining inaccessible to a large segment of India’s population.
He also pointed out that current evaluation benchmarks fail to capture “personal signatures” such as voice patterns, multilingual interactions, and contextual nuances that are critical for Indian users. According to Kumar, building indigenous datasets and evaluation standards tailored to India’s needs will be key to driving meaningful AI adoption across sectors like education, agriculture, governance, and healthcare.
His remarks come at a time when India is actively investing in domestic AI capabilities, supported by policy initiatives and growing private sector participation. The broader message from the summit was clear: for India to truly lead in AI innovation, it must create technology that speaks the language of its people — literally and culturally.
