Machine learning and artificial intelligence are becoming ubiquitous, but their impact on society are difficult to predict. There are fundamental issues regarding data handling and bias, value alignment, fairness and ethics in the design of machine learning algorithms that need to be taken into account to ensure there will be positive changes in the modern society. We study these issues in collaboration with social scientists, philosophers and computer scientists.
The data revolution is providing rich opportunities for the data science community to come together to identify novel solutions for using data about the state of economy and society to help governments, businesses and citizens make informed decisions for the future. It also raises important questions about data 'ethics, security and quality' as well as epistemic (leading to: unjustified actions, transparency, bias), normative (leading to: discrimination and challenges for autonomy; and informational privacy) and traceability (leading to moral responsibility) concerns around methods and algorithms used to gain insights from data and support the decision-making process.
The goal of this theme is to explore with various domain experts and policy makers the nature of these questions for the relevant applications (eg, medicine, criminolgy) and current responses to them and provide a tailored framework for the proposed methods and algorithms to 'act ethically' to strike a balance between controlling for various risks, while at the same time permitting 'optimal' complex data analysis.