Can code be responsible? - How to study and communicate algorithmic accountability

Abstract

Before we can trust algorithms, we must learn how to question and understand them. Based on this principle, we explore how algorithmic accountability can be taught through interdisciplinary, project- based learning as a part of the Digital Humanities Ruhr project. We present a seminar involving students of journalism, statistics, computer science, and data science. In interdisciplinary teams, they investigated real-world algorithmic systems and presented their findings in accessible public formats. The seminar demonstrates how algorithmic accountability can become an integral part of interdisciplinary education in the humanities and beyond. To accomplish this, we reflect on lessons learned and outline future improvements, such as expanding participation and enhancing mentoring to support both the selection of algorithmic cases and the choice of communicative formats. We integrate our improved seminar into the broader Data Literacy Certificate at TU Dortmund University, contributing to the development of micro-credentials in Digital Humanities across the University Alliance Ruhr.