Aarogya Setu 2.0: Revolutionizing Healthcare with Google’s Gemma 4 AI, ETHealthworld

June 30, 2026 Headlinesnews

New Delhi: The National Health Authority (NHA) has leveraged Google’s open AI model Gemma 4 and its Medical Data Toolkit in the recently launched Aarogya Setu 2.0 app.

The latest AI capabilities built into the app will enable users to process unstructured medical reports into comprehensive profiles of their individual health histories and health journeys.

The toolkit and Gemma 4 models are designed to identify the specific type of health record and extract critical information, including lab test names, methods, and results, from text and images.

It then maps and structures the information into standard Fast Healthcare Interoperability Resources (FHIR) data, a universal format that enables different healthcare systems to seamlessly share health records digitally, the tech giant said.

The integration supports the NHA’s efforts to enable smoother transitions of health information when patients move across hospitals and healthcare providers.

The move is part of the NHA’s and Google’s growing collaboration to strengthen India’s Digital Public Infrastructure (DPI) for healthcare, Google said.

The global tech giant also announced that it has open-sourced the Medical Data Toolkit for the benefit of the developer community and the broader digital health ecosystem.

“We are pleased that our AI capabilities are helping strengthen the country’s digital health stack and Digital Public Infrastructure. These tools are designed to empower Indians to more closely control their health journeys, while the release of the toolkit reduces barriers to entry for health tech innovation,” said Preeti Lobana, Vice President and Country Manager, Google India.

According to Google, the open-sourced Medical Data Toolkit can support developers and health organisations in digitising and standardising legacy health records while complying with interoperability standards under the Ayushman Bharat Digital Mission (ABDM). The initial version has been optimised for commonly used health records, including laboratory reports and observations.