Understanding User Perceptions of Trustworthiness in E-Recruitment Systems

Abstract

Algorithmic systems are increasingly deployed to make decisions that people used to make. Perceptions of these systems can significantly influence their adoption, yet, broadly speaking, users' understanding of the internal working of these systems is limited. To explore users' perceptions of algorithmic systems, we developed a prototype e-recruitment system called Algorithm Playground where we offer the users a look behind the scenes of such systems, and provide "how" and "why" explanations on how job applicants are ranked by their algorithms. Using an online study with 110 participants, we measured perceived fairness, transparency and trustworthiness of e-recruitment systems. Our results show that user understanding of the data Department: Head Editor: Name, xxxx@email Department Head 2IT professional and reasoning behind candidates' rankings and selection evoked some positive attitudes as participants rated our platform to be fairer, more reliable, transparent and trustworthy than the e-recruitment systems they have used in the past.

Improved Allocation and Reallocation Approaches for Software Trustworthiness Based on Mathematical Programming

“…Ogunniye et al developed a prototype e-recruitment system, to explore users' perceptions of algorithmic systems. They suggest that the data and reasoning behind the algorithmic results must be interpreted in order to improve users' perception of the fairness, reliability and transparency of e-recruitment systems, and to build trust in the systems [20]. Both references [21,22] use the fuzzy comprehensive evaluation method to assess software trustworthiness, but there is a difference in the methods for selecting trustworthy attribute sets and calculating trustworthy attribute weights.…”