As India advances toward an AI-enabled healthcare ecosystem, robust frameworks for data curation, sharing, and governance are essential to ensure that innovation is ethical, interoperable, privacy-preserving, and trusted. High-quality, discoverable, and reusable health data forms the foundation for responsible AI applications in healthcare, ranging from clinical decision support and population health management to medical research and personalized care.
The ICMR–National Institute for Research in Digital Health, in collaboration with the Koita Centre for Digital Health (KCDH), IIT Bombay, successfully hosted the Technical Working Group (TWG-1) Stakeholder Consultation Write-Shop on Data Curation, Sharing & Governance on 6–7 July 2026 at IIT Bombay.
ISPIRT participated in the event as subject matter experts (Tags – Dr Shyam Sundaram & Shikhar Verma), contributing perspectives on data empowerment, approaches to data Collaboration using data blinding framework , and the role of digital public infrastructure in enabling an AI-ready healthcare ecosystem in the context of the technical working group related to Data Curation, Sharing & Governance
The workshop brought together leading experts from government, academia, healthcare, industry, and policy domains to collaboratively develop recommendations for strengthening India’s health data ecosystem. Through focused and highly interactive discussions, participants examined the full health data lifecycle, including data curation, governance, interoperability, ethical and legal considerations, data sharing and reuse, AI model governance, operational frameworks, and capability maturity. The consultation provided a valuable platform for aligning stakeholder perspectives and advancing a shared vision for a secure, interoperable, and AI-ready digital health ecosystem for India.
In this context, ISPIRT highlighted as a feedback the importance of Data Empowerment and Protection Architecture (DEPA) as a foundational framework for empowering responsible AI enabling secure data sharing for AI model training.
