Biometrics systems are commonplace in the western world, and fraught with risk. For vulnerable populations, these risks of data privacy are potentially life threatening. In this session, we’ll explore the work that we’re doing integrating biometrics within healthcare and humanitarian programmes, as well as developing new Biometric Template Protection methods – ways of manipulating extracted biometric data to make them safer and add privacy-protecting properties – as a way to fundamentally shift the way that we deliver programmes and services in humanitarian contexts.
Facial recognition using AI models is currently used in many low-middle income countries as a way of linking people to their medical, financial and educational records when no official documentation exists. However, most models are biased towards lighter skin tones and are designed to run on high performance devices. At Simprints, we’ve built our algorithm of diverse and representative datasets, and our fingerprint scanner is 228% more accurate with scarred, worn, and damaged fingerprints (common in low resource settings).
We will also talk about how integrating biometrics within programme delivery has increased program efficiency (healthcare workers pulled up records 10x times faster than manual name search), and also acts as a way to increase transparency in development and ensure that every dollar, every vaccine and every service is reaching the intended recipient.
