CONTEXTUALIZING AI ETHICS IN TIME AND SPACE

Abstract

While the interest in AI ethics has overwhelmingly intensified over the last decade, and while various initiatives seek its institutionalization, the literature on algorithmic ethics tends to examine the subject through philosophical, legal, or technocratic perspectives, largely neglecting the empirical, socio-cultural ones. Moreover, this literature tends to focus on the United States, and to overlook other tech centers around the world. This paper aims to fill these gaps by focusing on how Israeli data scientists understand, interpret, and depict algorithmic ethics. Based on a pragmatist social analysis, and on 60 semi-structured interviews with Israeli data scientists, we ask: which ideologies, discourses and world views construct algorithmic ethics? And what cultural processes affect their creation and implementation? Our findings highlight three interrelated moral logics: A) ethics as a personal endeavor; B) ethics as hindering progress; and C) ethics as a commodity. We show that while data science is a nascent profession, these moral logics originate from the techno-libertarian culture of its parent-profession – engineering – and that they accordingly prevent the institutionalization of a wider, agreed-upon moral regime. We further show that these data scientists’ avoidance from institutionalized algorithmic ethics also stems from specific cultural and national determinants. Thus, this paper offers to see algorithmic ethics in a contextualized, culture-specific perspective, one that focuses on how data scientists practically see and construct their ethics, while considering their professional, organizational, and national contexts.