Scandinavian Journal of Information Systems
Exploring tensions in Responsible AI in practice. An interview study on AI practices in and for Swedish public organizations
The increasing use of Artificial Intelligence (AI) systems has sparked discussions regarding developing ethically responsible technology. Consequently, various organizations have released high-level AI ethics frameworks to assist in AI design. However, we still know too little about how AI ethics principles are perceived and work in practice, especially in public organizations. This study examines how AI practitioners perceive ethical issues in their work concerning AI design and how they interpret and put them into practice. We conducted an empirical study consisting of semi-structured qualitative interviews with AI practitioners working in or for public organizations. Taking the lens provided by the In-Action Ethics framework and previous studies on ethical tensions, we analyzed practitioners’ interpretations of AI ethics principles and their application in practice. We found tensions between practitioners’ interpretation of ethical principles in their work and ethos tensions. In this vein, we argue that understanding the different tensions that can occur in practice and how they are tackled is key to studying ethics in practice. Understanding how AI practitioners perceive and apply ethical principles is necessary for practical ethics to contribute toward an empirically grounded, Responsible AI.
Figueras, Clàudia; Verhagen, Harko; and Cerratto Pargman, Teresa
"Exploring tensions in Responsible AI in practice. An interview study on AI practices in and for Swedish public organizations,"
Scandinavian Journal of Information Systems: Vol. 34:
2, Article 6.
Available at: https://aisel.aisnet.org/sjis/vol34/iss2/6