Author: Miguel Sicart, Irina Shklovski, Mirabelle Jones
Publisher: N/A
Publication Year: 2021
Summary: The following paper attempts to answer a critical question that has vexed many debates: what constitutes an ethically accountable machine learning system? The authors of this paper investigate the ethical evaluation of machine learning methodologies. The authors examine machine learning techniques, methods, and technical practices through the lens of relational ethics, taking into account how machine learning systems interact with different forms of agency. Taking a cue from Phil Agre (1997), the authors employ the concept of a critical technical practice to analyze machine learning approaches. They argue that supervised learning appears to be the only ethically sound machine learning method.