Using data science and human-centered design to build responsible AI systems
I work on the key themes of uncovering and addressing idiosyncratic, as well as systemic risk arising due to lack of FATTER (Fair, Accountable, Transparent, Trustworthy, Explainable and Reproducible) AI. I have been involved in the Data for Social Good sector, having collaborated and worked with amazing organisations such as DataKind, Data4Change and Outsight International. My academic background is in Machine Learning, Informational Retrieval and Natural Language Processing. My work focuses on different aspects of an Equitable, Feminist AI Research: * DESIGN: methodologies and frameworks to help researchers, developers, data specialists and policymakers uncover challenges and opportunities in the application of AI * BUILD: tools and libraries to drive integration of responsible AI in the software systems * SCALE: tested methods and technologies * ADVOCATE: adoption of responsible data and AI frameworks
The issues pertaining to the responsible use of data and AI, or lack thereof, mostly arise from a unidimensional view - either from design, conceptualisation, implementation or monitoring - of AI systems. We need a holistic understanding of the evolutionary nature of data and AI and how they impact the systems, not just limited to software, that we interact with on a day-to-day basis. Through my work, I focus on the entire product lifecycle, from the inception all the way to the monitoring state, and different forms that intelligence and its by-products take during this lifecycle.