India floats AI plan under which ‘The model travels to the data, rather than the data travelling to the model’

India has proposed a new framework called the "Data Empowerment and Protection Architecture (DEPA) for AI Training." The concept is to train AI models on local data, rather than relying on offshore model-makers who may not prioritize the nation's issues or data owner control. The Ministry of Electronics and Information Technology (MeitY) is behind the scheme, which envisions "Training Data Providers" (TDPs) sharing anonymized and encrypted data with a "Data Exchange Platform" (DEP).

The TDPs include hospitals, banks, and government agencies. Entities called "Training Data Consumers" (TDCs) then access the data stored at the DEPs to train their models. TDPs can enforce rules on how TDCs use the data. The DEPA framework aims to prevent individual specific information from being learned or revealed by the model, using differential privacy techniques.

It also logs all activities inside secure data environments for audit purposes, ensuring security and transparency. Once implemented, the scheme could enable AI-driven diagnostics, drug discovery, public health management, secure sharing of financial data, personalized learning, and skill credentials. These advancements would help India realize its vision of inclusive growth and responsible AI leadership on the global stage.