Transparency by Design (Dagstuhl Seminar 21231)

Abstract

This report documents the program and outcomes of Dagstuhl Seminar 21231 on "Transparency by Design" held in June 2021. Despite extensive ongoing discussions surrounding fairness, accountability, and transparency in the context of ethical issues around AI systems that are having an increasing impact on society, the notion of transparency - closely linked to explainability and interpretability - has largely eluded systematic treatment within computer science to date. 
The purpose of this Dagstuhl Seminar was to initiate a debate around theoretical foundations and practical methodologies around transparency in data-driven AI systems, with the overall aim of laying the foundations for a "transparency by design" framework – a framework for systems development methodology that integrates transparency in all stages of the software development process. Addressing this long-term challenge requires bringing together researchers from Artificial Intelligence, Human-Computer Interaction, and Software Engineering, as well as ethics specialists from the humanities and social sciences, which was a key objective for the four-day seminar conducted online.