Ethics, Integrity and Good Practice in ML

Speaker: Ross Clarke, University of Cambridge Abstract: Thanks to an explosion of popularity and technological success, machine learning and artificial intelligence in their various guises now touch nearly every aspect of our lives. Every day, hundreds of new papers appear promising to revolutionise yet another facet of the field. But how often do we consider the broader impact of our work? Is it appropriate for certain lines of recent research to exist in their present forms? And if general intelligence isn’t yet just around the corner, do we even need to be asking these questions? In this talk, we give an overview of current work investigating the ethics of machine learning and its applications. We explore examples of harms resulting from real-world ML deployments, discuss what it means to be an ‘ethical’ researcher and survey views from the literature of how the field might become more ethics-focussed. Discussion and participation strongly encouraged! No pre-reading is required. We draw from several references, which will be made available with our slides for later review.