Data Ethics and Ethical AI: Key Issues and Approaches
The recent transition of AI from a lab-based science to its deployment in live human contexts brought into sharp focus a plethora of historic socio-cultural biases, inequalities and moral dilemmas. Although, we have access to a growing number of increasingly concrete examples of how data and algorithmic bias may affect (for better and worse) individuals, specific groups and society more generally, and despite numerous proposals of high-level ethical principles that should guide both data sciences and AI engineering and deployment, our knowledge of how to action those principles in technology designs remains limited. This is largely owed to the complex nature of ethics itself and to our lack of understanding of how our human biological brains and socio-cultural minds interact with technology. In this lecture, I will set the scene for developing such an understanding by reviewing the key issues and approaches that relate to data and AI bias, the harms that such biases often entail and how data and AI scientists are beginning to engage with the idea of the data and AI ethics by design. Given my own focus on the application of AI in education, I will demonstrate these issue and approaches through examples which are predominantly based in education – a particularly complex domain of application that entails profound societal and economic consequences if mishandled.
Biographical note
Kaśka Porayska-Pomsta is Professor of Artificial Intelligence in Education at UCL. Her research focuses on developing AI systems for education and evaluating fundamental questions about how AI can be designed to support human learning and development. Her research, which bridges between engineering and social sciences disciplines, is embedded in front-line education and other in-the-wild contexts. She has 20+ years of experience of successfully leading complex, highly ambitious, inter-institutional projects aiming for real world impact worth over £15M. Kaśka actively contributes to UK policy on AI and Ethics. She is co-editor of special issue of the International Journal of AI in Education on Fairness, Accountability, Transparency and Ethics of AI in Education, and co-editor of a forthcoming book on Ethics of AIED. She recently completed her 3-year service as Departmental Head of Research at the UCL Institute of Education, she is member of the management committee for the Centre for Educational Neuroscience, and member of the Executive Board and a co-chair of the Inclusion working group of the International Society for AI in Education.