Hiring Algorithms: An Ethnography of Fairness in Practice

Although AI is claimed to be superior over human experts in terms of objectivity, efficiency, and effectiveness (Hipps 2017; Miller 2015), its growing influence on the outcomes of our daily lives has become increasingly a source of concern. Critics have accused AI of committing ethical violations, such as being racist, sexist, or harming privacy (Crawford and Calo 2016; O'Neil 2016). For example, Amazon recently had to withdraw its automatic hiring algorithm after discovering that it consistently downgraded resumes of female candidates (Dastin 2018). Several agendas have therefore emerged aimed at addressing and solving ethical issues surrounding AI applications, for example, by calling for regulations to enforce transparency, simplifying machine learning models, or devising auditing algorithms that can evaluate discriminatory outcomes (Crawford and Calo 2016; O'Neil 2016; Pasquale 2015; Selbst et al. 2019). However, so far literature offers little insight into how organizations deal with ethical values and AI in practice. Therefore, this research aims to address the following research question: how do organizational groups engage with AI in practice and how are their ethical values being reconfigured?