Trust in Artificial Intelligence: Comparing trust processes between human and automated trustees in light of unfair bias

Abstract

Introducing automated systems based on artificial intelligence and machine learning for ethically sensitive decision tasks requires investigating of trust processes in relation to such tasks. In an example of such a task (personnel selection), this study investigates trustworthiness, trust, and reliance in light of a trust violation relating to ethical standards and a trust repair intervention. Specifically, participants evaluated applicant preselection outcomes by either a human or an automated system across twelve personnel selection tasks. We additionally varied information regarding imperfection of the human and automated system. In task rounds five through eight, the preselected applicants were predominantly male, thus constituting a trust violation due to a violation of ethical standards. Before task round nine, participants received an excuse for the biased preselection (i.e., a trust repair intervention). Results showed that participants initially perceived automated systems to be less trustworthy, and had less intention to trust automated systems. Specifically, participants perceived systems to be less able, and flexible, but also less biased – a result that was sustained even in light of unfair bias. Furthermore, in regard to the automated system the trust violation and the trust repair intervention had weaker effects. Those effects were partly stronger when highlighting imperfection for the automated system. We conclude that it is crucial to investigate trust processes in relation to automated systems in ethically sensitive domains such as personnel selection as insights from classical areas of automation might not translate to application contexts where ethical standards are central to trust processes.

Ähnliche Arbeiten