Gaining transparency in Uber’s algorithmic management

This empirical study sheds light on Uber’s algorithmic management and provides the first calculation of working time with Uber drivers in Paris to understand the way digital platforms transform the temporalities of work. Based on the personal data of 12 Parisian Uber drivers, we show how Uber organizes drivers’ work through the allocation of trips and calculation of distances according to three distinct working-time units: “searching for, approaching, and driving with passengers”. The data analyses show that drivers are increasingly working hours and driving kilometers for the platform when searching for passengers, which is unpaid work but crucial for algorithmic operations. In addition, the significant variability in working hours and rates generated depends on the behavior of the market, which is invisible to drivers. Nevertheless, this dependence and the resulting economic losses and gains are passed on to drivers individually as “algorithmic predictability risk”. Indeed, drivers’ data are controlled by Uber to make automated predictions of supply and demand for the benefit of the platform exclusively.