Particles of a Whole: Design Patterns for Transparent and Auditable AI-Systems
|Published in:||Adjunct Proceedings of the 2021 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2021 ACM International Symposium on Wearable Comput- ers (UbiComp-ISWC ’21 Adjunct), September 21–26, 2021, Virtual, USA, 272-275|
The integration of principles such as transparency and auditability in AI-systems is a challenging endeavour. The demand for guiding concepts is to be not too high-level and not too specific at the same time. In this paper, I will introduce design patterns as a promising method to meet this demand and as a guidance for the implementation of transparency and auditability in sociotechnical AI-systems. This will include the history of design patterns in urban planning and software engineering, as well as their distinguishing features. In extension to this, I outline a pattern approach for responsible AI-systems in a democratic society. This includes a sketch of possible design patterns for transparent and auditable AI-systems. Some early challenges and open questions are discussed at the end, together with the conclusion.
Connected HIIG researchers
- Open Access