Abstract

The use of algorithmic systems in digital platforms that aid humans in task performance and decision-making is becoming increasingly prevalent. However, the black-box nature of algorithms has led to critical issues regarding the transparency of these systems. An illustration of this can be seen through the experiences of Uber drivers, who feel uncertain about how the Uber algorithm works, how the rides are allocated, and how their earnings are calculated (Möhlmann et al. 2021). This is detrimental not only to the platform employees and users themselves but also to the company as a whole. Several digital platforms such as Uber have been sued and even fined due to this particular issue. However, the focus of transparency discussions has primarily been on technical aspects and insufficient attention has been given to individuals as recipients of transparency. Given that transparency is a subjective concept that is shaped based on individuals’ mindset (Shin and Park 2019), it is crucial to comprehend the factors that influence their perception of transparency, rather than solely relying on technical aspects. In this study, we will shift the focus toward the recipients of transparency and enhance our understanding of the factors that shape their perception of algorithmic transparency (AT). In line with this argument, we know that individuals do not form their perception based on one isolated factor but multiple factors simultaneously work together to form one’s perception of AT. Therefore, we aim to review IS and information transparency literature, identify factors that influence AT from the perspective of digital platform workers (whose works are managed by algorithm) and digital platform users (who use algorithmic information for their decision-making), and investigate how the configuration of such factors influences their perception. To meet this end, we will conduct a survey to collect data from the recipients of transparency in Uber and Lyft platforms. The selection of these two platforms was based on their high score concerning their algorithmic mechanism. We will test our research model using fuzzy-set Qualitative Comparative Analysis (fsQCA) using the QCA software program (Ragin 2000). fsQCA is a suitable approach to test the model because it utilizes set memberships to identify configurations of conditions influencing an outcome, making it ideal for exploring the relationship between combinatory effects of antecedents and individuals’ perception of AT. Our study makes important contributions to IS research through a more nuanced conceptual understanding of antecedents of transparency perception in algorithmic systems. Also, this article contributes to transparency studies by focusing on the recipients of transparency and offering insights into the factors that shape their perception of AT. From a practical standpoint, this article helps organizations mitigate transparency-related issues by understanding why different individuals have varying interpretations of transparency.

Download

Share

COinS