Hard choices in artificial intelligence

• DOI: 10.1016/j.artint.2021.103555
• Publication Date: 2021-07-14T15:19:13Z
• Authors: Roel Dobbe; Thomas Krendl Gil...; Yonatan Mintz
• Keywords: Sociotechnical system; Vagueness; Deliberation

As AI systems are integrated into high stakes social domains, researchers now examine how to design and operate them in a safe ethical manner. However, the criteria for identifying diagnosing safety risks complex contexts remain unclear contested. In this paper, we vagueness debates about behavior of systems. We show cannot be resolved through mathematical formalism alone, instead requiring deliberation politics development as well context deployment. Drawing from new sociotechnical lexicon, redefine terms distinct challenges at key stages system development. The resulting framework Hard Choices Artificial Intelligence (HCAI) empowers developers by 1) points overlap between decisions major challenges; 2) motivating creation stakeholder feedback channels so that issues can exhaustively addressed. such, HCAI contributes timely debate status democratic societies, arguing should goal Safety, not just procedure which it is ensured.