As India steps up efforts to develop sovereign artificial intelligence systems, the country must focus on creating culturally rooted AI by making government-held data widely available, EY India’s Rajnish Gupta has said.

Ahead of the Union Budget, Gupta said most global Large Language Models are heavily influenced by western data, particularly from the US and Europe, leading to responses that often fail to reflect India’s social, cultural and linguistic realities. He warned that without access to large volumes of Indian data, local developers would struggle to build AI systems that truly represent the country’s diversity.

Gupta, Partner in EY India’s Tax and Economic Policy Group, noted that India’s complex mix of languages, traditions and religious nuances is frequently misunderstood or ignored by existing AI models. As a result, answers generated by such systems tend to adopt western perspectives, limiting their relevance for Indian users.

He stressed that the government should play a key role by acting as a primary data provider. According to him, vast amounts of anonymised public data can be responsibly released to help developers train AI models that are more accurate, inclusive and culturally aligned with India.

Comparing AI development with India’s successful digital public infrastructure initiatives such as Aadhaar and UPI, Gupta said the country has already shown how technology can be democratised at scale. He suggested that treating data and computing capacity as public goods could lay a strong foundation for India’s AI ecosystem.

While advocating public investment in data infrastructure, Gupta cautioned against overregulation. He welcomed the government’s decision not to mirror the European Union’s strict AI laws, arguing that excessive controls could hinder innovation.

Gupta urged policymakers to maintain a hands-off approach, avoid licensing barriers, new taxes or regulatory friction, and allow Indian entrepreneurs the freedom to experiment. He expressed confidence that with the right environment, India can develop globally competitive AI systems that are deeply rooted in its own culture and values.