In recent years, labour platforms have gained great visibility and prominence in labour studies and public debate. In particular, data work platforms allow large projects to be broken down into small tasks and assigned to anonymous workers, each of whom remotely executes a tiny part of the project and receives compensation for it.
Empirical evidence consistently shows that these workers do not know other people using these platforms, that they talk little about this work with their family and friends, and that in general, they do not know anyone else who is doing these same tasks. These experiences contrast with the findings of studies looking at other platform-based work. We propose here to wonder about social ties as a question before that of collective organisation and, therefore, of any instance of regulatory discussion or contestation action on the part of workers.
In this paper, we address two issues relevant to understanding the work experience of data workers. On the one hand, we analyse how these workers arrive at data task platforms. This search is fundamentally digital, with little relevance of personal networks. In particular, weak ties play an insignificant role: for those who arrived on the recommendation of someone, it was mainly family or very close friends, not networks of colleagues or ex-workmates. On the other hand, we analyse the workspace in which these tasks take place, both virtual - the platform - and physical - usually their home - and contrast this with findings from other studies on physical work platforms and other kinds of online work platforms. The contrast between platforms provides relevant clues to understand which arrangements enable some kind of bonding and solidarity between workers, and which ones hinder or block it.
The analysis is based on three types of sources. Firstly, a survey of 220 workers based in Argentina using the Microworkers (between December 2020 and February 2021) and Clickworker (between March and June 2022) platforms, carried out in the framework of the scientific project TRIA (TRabajo de la Inteligencia Artificial). Secondly, semi-structured interviews were conducted with a selected group of participants. Finally, we extracted from the interviews names of social media channels and groups where this form of work is discussed, we collected, and carried out content analysis of posts, videos and tutorials published there.