Total: 1
Ride-sharing platforms like Uber market themselves as enabling `flexibility' for their workforce, meaning that drivers are expected to anticipate when and where the algorithm will allocate them jobs, and how well remunerated those jobs will be. In this work we describe our process of participatory action research with drivers and trade union organisers, culminating in a participatory audit of Uber's algorithmic pay and work allocation, before and after the introduction of dynamic pricing. Through longitudinal analysis of 1.5 million trips from 258 drivers in the UK, we find that after dynamic pricing, pay has decreased, Uber's cut has increased, job allocation and pay is less predictable, inequality between drivers is increased, and drivers spend more time waiting for jobs. In addition to these findings, we provide methodological and theoretical contributions to algorithm auditing, gig work, and the emerging practice of worker data science.