With the advent of Artificial Intelligence (AI) in both public health and healthcare, the digital divide is seen to be growing even more substantial in Low- and Middle-Income Countries (LMICs) in comparison to High-Income Countries (HICs). The reasons for this foreseen development are various: a limited number of large companies are dominating the market, training models particularly utilizing English language documents for training data neglecting languages leading to bias, non-affordability of smart phones hold market and lack of training data utilizing diverse cultural information lead to bias in algorithm development.
This session will address these issues, thriving to discuss possible barriers and solutions to circumvent these barriers to narrow the digital gap particularly caused by the role-out of AI based medical applications, focusing on global public health. Issues that should be tackled are contexts for measurement (identify AI components and functions within operational system, outputs and risks, service context and boundaries, identifying key stakeholders and monitoring goals (e.g. safety, usability, transparency, reliability, patient safety etc.), as well as methods for indicator selection and categorization and feedback from different user groups.
