India Proposes Decentralized AI Framework to Secure Local Data

India's Ministry of Electronics and Information Technology (MeitY) has introduced a framework called the “Data Empowerment and Protection Architecture (DEPA) for AI Training” to ensure local data remains under domestic control while powering artificial intelligence. According to reporting from The Register, the initiative is driven by the government's belief that offshore model-makers lack the motivation to address India's specific national issues.

Professor Ajay Kumar Sood, Principal Scientific Adviser to the Government of India, describes the DEPA approach as one where “the model travels to the data, rather than the data travelling to the model.” Under this system, “Training Data Providers” (TDPs)—such as government agencies, banks, and hospitals—share encrypted and anonymized data with a “Data Exchange Platform” (DEP). “Training Data Consumers” (TDCs) then access this data within “hardware-based Secure Data Environments” to train their models. MeitY notes that these environments keep data isolated and encrypted during processing, utilizing differential privacy techniques to prevent the disclosure of individual information.

MeitY envisions several high-impact applications for this deployment. In healthcare, the goal is to integrate fragmented data for drug discovery and AI-driven diagnostics. In the financial sector, the government aims to extend formal credit to 400 million underserved individuals and MSMEs. Additionally, the framework targets India's $29 billion EdTech sector to provide personalized learning for 250 million students.

This move aligns with India's broader digital public infrastructure (DPI) strategy, which includes tools like the Bharat Taxi ride-sharing platform and the Unified Payments Interface to prevent Big Tech dominance. The DEPA framework positions this approach as a contribution to global science diplomacy, supporting efforts by the United Nations and the World Bank to create a more inclusive AI ecosystem.