In AI, India’s choice needn’t be between a US leash or Chinese hook

The article highlights India's precarious position regarding foreign AI models. The temporary cessation of access to US-developed AI models demonstrated how Washington could disrupt India's AI infrastructure. While Chinese open-source models offer an alternative that cannot be switched off remotely once downloaded, they reportedly come with built-in political sensitivities, as illustrated by a chatbot's refusal to answer a question about a disputed Indian state.

This situation forces India to choose between two imperfect options: US models that can be centrally controlled, or Chinese models that may carry embedded geopolitical biases. The piece suggests that neither path is ideal for a nation of 1.4 billion people, particularly for critical services like banking, courts, and public administration, which require unfettered and neutral AI operation.

India's current response involves significant domestic investment, with initiatives like Sarvam AI and BharatGen planning to develop large-scale AI models. The government is also exploring substantial anchor investments of up to 200 billion rupees (US$2.08 billion) in a new frontier AI fund, building on the 103.7 billion rupees approved for the "IndiaAI Mission" in 2024.

The core argument is that India's priority should be the ability to adapt any AI model, regardless of its origin, to serve specific Indian needs, ensure its functionality, and maintain its operation independently. This emphasizes the importance of customization and operational sovereignty over mere access to foreign-developed technologies.