Perception of Fairness in Algorithmic Decisions: Future Developers' Perspective

While professionals are increasingly relying on algorithmic systems for making a decision, some occasions, decisions may be perceived as biased or not just.  Prior work has looked into the perception of decision-making from user's point view. In this work, we investigate how students in fields adjacent to algorithm development perceive FATE decision-making. Participants (N=99) were asked rate their agreement with statements regarding six constructs that related facets fairness and justice three separate scenarios. In addition, they to: a) define fairness, b) discuss possible causes unfairness system, c) suggest methods increasing transparency d) provide view accountability case system behaves unfairly. The findings indicate 1) 'agreeing' decision does mean person 'deserves outcome', 2) perceiving factors used 'appropriate' make 'fair' 3) system's 'not fair' is affecting participants' 'trust' system. In participants found proportional distribution benefits more fair than other approaches. Furthermore, level academic education role play decision-making. qualitative part study revealed most likely defined use 'objective factors' identify 'sensitive attributes' cause unfairness. Finally, adding 'explanations process output' preferred strategy increased if unfairly, 'development team' should held accountable.