AI Facial Recognition Is Denying Food To Pregnant Women Across India

A Decode investigation reveals how the use of AI-driven facial recognition in India’s Integrated Child Development Services (ICDS) is denying food to some of the country’s most vulnerable women and children.

Since July 2025, pregnant women and nursing mothers must authenticate their identity through a facial scan on the government’s Poshan Tracker app to receive nutrition benefits. However, the system frequently fails to match live images with years-old Aadhaar photographs, especially as pregnancy, ageing, and health conditions alter facial features.

The report finds that nearly half of eligible beneficiaries have been unable to access rations due to authentication failures, raising serious concerns about exclusion in a rights-based welfare programme. It also uncovers that the system relies on Google’s ML Kit, a tool designed for face detection rather than long-term identification, which was not publicly disclosed by the government.

Through on-ground reporting across multiple states, the story highlights how technological gaps, poor infrastructure, and rigid verification requirements are reshaping welfare delivery, placing the burden of failure on both right-holders and frontline workers.