As India moves rapidly towards building sovereign artificial intelligence capabilities, the country must prioritise the public release of government-held data to counter the Western bias embedded in most global AI models. Speaking ahead of the Budget, AI systems that truly reflect India’s cultural, linguistic, and social diversity. The most widely used Large Language Models (LLMs) today are largely trained on datasets from the United States and Europe, leading to outputs that often overlook or misunderstand Indian contexts.
India’s cultural nuances, multilingual landscape, and religious complexities are insufficiently represented in existing global models. To address this gap, it is important for the government to act as a primary data provider by making public datasets more accessible to domestic AI developers.
Drawing parallels with landmark digital public infrastructure initiatives like Aadhaar and the Unified Payments Interface (UPI), India could similarly democratise AI by treating data and computing infrastructure as public goods. While public investment should not dictate innovation, it plays a vital role in laying a strong foundation for private-sector experimentation and growth.
At the same time, the government must refrain from adopting regulatory frameworks similar to the European Union’s AI laws, arguing that minimal interference is essential to foster innovation. It is essential to avoid excessive licensing requirements; new taxes and regulatory friction would prevent Indian entrepreneurs from developing globally competitive, culturally rooted AI models from within the country.