India Seeks AI Strategic Autonomy to Curb Data Transfer Risks
MeitY Secretary S Krishnan highlights risks of proprietary AI models and India's push for domestic AI stack and strategic autonomy.
The Ministry of Electronics and Information Technology (MeitY) has raised concerns regarding the use of proprietary artificial intelligence (AI) models, warning that they pose significant risks to national data security. Speaking on the sidelines of the IndiaAI Mission and the World Bank Group's launch of the World Development Report 2026, MeitY Secretary S Krishnan emphasized that India is actively working to build its own domestic AI ecosystem to preserve strategic autonomy.
The Risk of Proprietary AI Models
A primary concern for the government is the potential for Indian data to be transferred outside the country when using foreign-owned, proprietary AI systems. Beyond the movement of information, there is a deeper technical risk: these models are designed to learn from the information provided to them.
"Where proprietary models are to be used, we need to be aware that it involves the risk of data transfer and it involves the risk of those models learning at the cost of our data and at the cost of getting trained on what we give them," Krishnan explained to reporters.
Building a Domestic AI Stack
To counter these risks and reduce dependence on external technology providers, India is focusing on developing a comprehensive, homegrown AI technology stack. This massive undertaking aims to cover every layer of the technology, including:
- Physical Infrastructure: Energy systems, specialized chips, and servers.
- Computing Capacity: The raw power required to run complex models.
- Data and Models: The foundational information and the algorithms themselves.
- Applications: The end-user software used by citizens and industries.
A Balanced Approach: Open Source vs. Proprietary
While the goal is domestic capability, the government is not looking to isolate itself. Krishnan noted that India is exploring the use of open-source and open-weight AI models. These models offer a significant advantage because they can be deployed locally without the need to transfer sensitive data out of the country.
However, the government intends to maintain a "judicious mix" of technologies. "We are an open country both economically and socially," Krishnan stated, noting that the objective is to use a combination of measures to ensure the nation's overall strategic autonomy remains intact while expanding AI adoption across the country.
Data Localization and Governance
When asked if companies and startups empanelled by the government would be mandated to keep data within India, Krishnan clarified that the requirements would depend on the specific nature of the data. He pointed to existing government classifications for different data categories, which dictate the necessary standards for data centres and cloud services.
As India seeks to integrate AI into government operations and various industries, the tension between rapid adoption and data sovereignty remains a central pillar of the nation's technological roadmap.