Carl Westin

4 Source bias effects There appears to be differences in how human and automated de- cision aids are judged, and these differences are intertwined with the perceived expertise of the source. For example, errors made by automated sources tend to affect reliance more negatively than errors made by human sources. Possibly, such biases may have at- tributed to the rejection of advisories, particularly conformal ones, in the first empirical study described in the previous chapter. As such, this chapter aims to investigate, both through literature and a real-time study, whether those rejections were tied to a general bias against automation.