India's Language Diversity and AI: Who Decides Which Languages Go Digital

India is home to an extraordinary linguistic diversity — hundreds of languages are spoken by millions of people. Yet as artificial intelligence reshapes communication and access to information, most of these languages face a fundamental challenge: they lack the digital data AI needs to learn them.

The Data Problem

Large language models are trained predominantly on text from the internet, which is overwhelmingly dominated by English, Mandarin, and a handful of other major languages. For India's smaller languages — many with rich literary traditions but limited online presence — this creates a visibility gap. Without sufficient training data, AI systems simply cannot process or generate text in these languages effectively.

AI as Potential Equalizer

Some researchers and startups see an opportunity: by creating datasets for underrepresented Indian languages and training specialized models, AI could become a tool for linguistic preservation and promotion. Applications range from voice assistants in regional languages to translation tools that let rural communities access government services.

However, critics warn that the power to determine which languages receive investment and data resources lies largely with major tech companies — raising questions about whether commercial incentives align with cultural preservation goals. Languages with the largest speaker populations are most likely to attract corporate interest, potentially widening the digital gap for truly minority tongues.