“…ADS is often evaluated according to to specific normative criteria, including fairness [14,71], accountability [147], explainability [8,103], and contestability [67,140]. Such inquiry is motivated in part by evidence that ADS can be biased and can cause various types of harms [13,15,31,109], and in part by application of theories of justice [18,75,98]. Despite major theoretical and practical issues commonly known in academia, prior work shows that perception of the trustworthiness of algorithms relative to humans can be favorable, though it is highly dependent on context, subjectivity of the domain, and performance [11,89,95].…”