Paper short abstract:

This paper, based on the study of the two-sided rating system used by Uber, will show how anthropology and computational methods can be combined to better understand the perceptions of the riders and the drivers who rate each other after every trip.

Paper long abstract:

As Uber's stake in India, particularly in the metros, is on the rise, it becomes important for this ridesharing company to maintain the quality of service, trust, and reliability when moving people around the city space. To ensure this while also maintaining employee satisfaction, Uber tends to follow the two-sided rating system where both the rider and the driver judge and rate each other after every trip.

Though popularly advertised by Uber as a trust-building exercise to increase cooperation and respect between the rider and the driver, such rating through algorithmic design also has an impact on the driver-rider matching when ordering a cab service, thus in a way affecting the riding experience and the value for money. The system has several unintended consequences in a socio-cultural context such as India and very specifically for the middle class who use this service often. The pressure to maintain a certain high score very often gets juxtaposed with the burden to present themselves and constantly judge each other in ways that sometimes increase rater bias.

This paper using both ethnography from an anthropological perspective and computational methods intends to show how the riders and the drivers perceive and rate each other and how such perceptions may lead to rating collusions and sometimes in collisions.