MeitY Launches DEPA Handbook to Supercharge AI Training
MeitY introduces the DEPA handbook to enable secure, privacy-protected data sharing for AI training, ensuring institutions retain control of their data.
India's quest to become a global leader in artificial intelligence has taken a significant leap forward. The Union Ministry of Electronics and Information Technology (MeitY) has officially launched the DEPA handbook for AI training, a strategic framework designed to unlock massive amounts of high-quality data while keeping privacy and security at the forefront.
Unveiled during a high-level roundtable in Bengaluru by S. Krishnan, Secretary of MeitY, and supported by the iSPIRT Foundation, the Handbook on Data Empowerment and Protection Architecture (DEPA) aims to dismantle the barriers that have long hindered data sharing between institutions.
Solving the Data Dilemma: Privacy vs. Progress
For AI to evolve, it needs vast amounts of data. However, institutions often hesitate to share this information due to fears regarding intellectual property, confidentiality, and loss of institutional control. The DEPA framework addresses these pain points by establishing a "techno-legal" structure for purpose-bound collaboration.
Instead of a free-for-all, the framework employs a sophisticated blend of:
- Contractual Permissions: Ensuring data is used only for agreed-upon purposes.
- Controlled Computing Environments: Creating secure spaces where data can be processed without risk of leakage.
- Privacy-Enhancing Safeguards: Implementing technical barriers to protect sensitive information.
- Data Provenance: A tracking system that monitors the origin, permissions, and usage of data throughout the AI training lifecycle to ensure total transparency and accountability.
"India's AI ambitions must be matched by the ability to use data responsibly and securely. The DEPA Training Framework enables trusted data collaboration at scale, unlocking access to large-scale datasets for AI training while safeguarding privacy, commercial confidentiality and institutional control," stated S. Krishnan.
A Revolutionary Shift: The Model Travels to the Data
One of the most striking features of this framework is the reversal of how data is typically handled. Ajay Kumar Sood, the Principal Scientific Adviser to the Government of India, explained a paradigm shift in data movement.
Traditionally, data is transferred to the AI model for training. Under the DEPA framework, the model travels to the data. This means institutions do not have to surrender or transfer their underlying information; the AI learns from the data within the institution's own secure environment.
This approach is expected to:
- Facilitate multi-party and cross-border data collaborations.
- Accelerate the pace of scientific discovery.
- Position India as a key player in global science diplomacy by promoting a responsible and inclusive AI ecosystem.
The 'UPI Moment' for India's Data Economy
The significance of this launch was highlighted by industry veterans who see it as a foundational shift for the nation's digital infrastructure. Sharad Sharma, Co-founder of the iSPIRT Foundation, compared the launch of the DEPA framework to the introduction of the Unified Payment Interface (UPI) in 2016, suggesting it could be equally transformative for the data economy.
Adding to this, Kris Gopalakrishnan, Co-founder of Infosys and Chairman of Axilor Ventures, noted that the framework empowers researchers to build more robust and diverse AI models without compromising the rights of the data contributors or custodians.
The Bengaluru roundtable served as a melting pot of ideas, bringing together leaders from academia, government, industry, and research to discuss how trusted data collaboration can be leveraged for the public good and rapid AI innovation.