.jpg)
MeitY Launches DEPA Handbook for AI Training to Promote Secure Data Collaboration
Bengaluru: In an effort to promote secure data collaboration for Artificial Intelligence (AI) development, the Ministry of Electronics and Information Technology (MeitY), Government of India, has launched the Handbook on Data Empowerment and Protection Architecture (DEPA) for AI Training. The handbook was unveiled by S. Krishnan, Secretary, MeitY, in a high-level roundtable in Bengaluru on October 31, 2026, in collaboration with the iSPIRT Foundation.
One of the biggest challenges with AI development is having access to large volumes of diverse and high quality data. Whilst this can help improve research and training for AI development, organisations may be reluctant to share their data due to privacy, commercial sensitivity, IP and other concerns. The DEPA Training framework aims to allow data collaboration without relinquishing control.
What is DEPA for AI Training?
DEPA stands for Data Empowerment and Protection Architecture. It provides a techno-legal framework for data collaboration with specific purposes, such as when multiple institutions need to work on data sets. The four main components of the framework are:
- Data Sharing Agreements: Define what purpose the data can be used for.
- Secure Computing Environments: Use the data in a secure environment.
- Confidentiality: Confidentiality measures to address privacy, sensitivity and control.
- Audit Trail: Provide a trail of where the data came from, what permissions were given and how it was used for training.
This will provide a secure and transparent system to enable institutions to work on AI.
How will DEPA Help with Secure Data Collaboration?
The framework aims to enable data collaboration but reduce the need for institutions to relinquish control. "The whole idea is that instead of the data going to the AI model, we bring the AI model to the data," explained Prof. Ajay Kumar Sood, Principal Scientific Adviser to the Government of India, at the launch.
This can help organisations to work with data sets but give them greater control over where their data is stored. This is particularly useful for research and development projects that require multiple institutions to work together. Bringing different data sets together will help solve complex problems. The framework also provides an audit trail to help understand where the data came from and what permissions were given. It provides a way for organisations to work with data securely and transparently.
MeitY Secretary S. Krishnan Emphasises Responsible AI Development
"We need to ensure that our ambitions in the AI space are achieved responsibly and securely using data," said MeitY Secretary S. Krishnan. "It has the potential to enable greater access to large volumes of data for training AI whilst keeping data private, commercially sensitive and providing control. We are committed to moving this framework forward to enable innovation in AI here in India," he added. Attendees from the government, industry, technology, academia, research and media industries discussed how data collaboration can be used to innovate in AI.
Industry and Research Leaders Highlight Benefits
Sharad Sharma, Co-founder iSPIRT Foundation said this was a huge milestone for India’s data economy. "It is as big as the announcement of UPI in February 2016." Anjani Rathor, Group Head – Digital Banking and Customer Experience, HDFC Bank, said that data collaboration can help in the banking industry. "We have used DEPA Inference and this gives us an opportunity to build on this with DEPA Training."
Kris Gopalakrishnan, Co-founder Infosys and Chairman of Axilor Ventures said that having secure access to diverse data sets can help researchers improve their AI models and collaboration between teams. It also ensures the interests of data holders. Hari Subramanian presented on the DEPA Training framework and there were contributions from Emergent AI, Sarvam and YourStory.
What Does DEPA Mean for India’s AI Ecosystem?
Having access to accurate data is a key part of developing and testing AI systems. However, organisations may be reluctant to share sensitive or valuable data which can limit research collaboration. By providing a framework that works with data sharing agreements, secure computing environments and audit trails, DEPA Training aims to provide a structure for collaboration. This can be used across multiple industries including banking, research, technology and more.
It also puts control back into the hands of the institutions. This is important as it doesn’t mean everyone will have unrestricted access to the data. This launch is a step towards creating a secure data sharing ecosystem for AI development in India. How it is implemented will determine how impactful it is.Overall the message from MeitY is that to move AI forward we need not only powerful models but also secure ways to work with data between organisations.
FAQs
Q1. What is the DEPA for AI Training Handbook?
The DEPA for AI Training Handbook, launched by MeitY, enables purpose-bound data collaboration for artificial intelligence training, and enables institutions to control their data.
Q2. What is DEPA?
DEPA is an acronym for Data Empowerment and Protection Architecture.
Q3. When was the DEPA for AI Training Handbook launched?
MeitY launched the handbook on October 31, 2026, during a high-level roundtable in Bengaluru.
Q4. Who released the DEPA for AI Training Handbook?
The handbook was released by S. Krishnan, secretary, Ministry of Electronics and Information Technology, in the presence of the iSPIRT Foundation.
Q5. What are the main elements of the DEPA Training framework?
The framework comprises trust-based data collaboration enabled by contractual authorization, controlled computing environments, privacy-enhancing safeguards and data provenance.
Q6. In what way does DEPA protect data during AI training?
It allows institutions to collaborate on data for AI training and track the permissions and origins of the data during the process.
Click Here for More Science & Technology