Apulaisprofessori, dosentti Hiippala Tuomo ja työryhmä

Digital Sweatshops as Research Infrastructure? Critical and Practical Perspectives to Crowdsourcing in Academic Research

Tieteellinen tutkimus / siihen pohjautuva työ | Nelivuotinen

This project examines crowdsourcing as a part of the academic research infrastructure. In this context, crowdsourcing refers to privately-owned online platforms where workers perform small and repetitive tasks for money. In addition to powering voice-controlled assistants, self-driving cars and keeping our social media platforms clean of abusive content, crowdsourcing is widely used in academia. Computer vision, natural language processing and other data-intensive fields depend on crowdsourcing to create data for training and evaluating algorithms, but the use of crowdsourcing is now also expanding to social sciences and humanities. This warrants urgent attention, because crowdsourcing involves various ethical issues ranging from sweatshop wages to diverse forms of invisible labour. In this project, we seek to understand the role of privately-owned crowdsourcing platforms in academic research through three objectives. We aim to (1) understand how privately-owned crowdsourcing platforms control the labour and constrain the agencies of workers and requesters, (2) develop methods and tools for fair and ethical use of crowdsourcing platforms, and (3) reconceptualise academic crowdsourcing as a form of participatory research governed by fair and transparent algorithms. To achieve these objectives, we draw on a novel combination of sociological perspectives to platforms and socio-technical systems with approaches from the fields of digital humanities and human computation. We conduct interviews and surveys with crowdsourced workers and stakeholders, and combine methods for qualitative and automatic text analysis to study online discussions and public documentation related to crowdsourcing platforms. We also develop computational tools and methods for ethically-responsible crowdsourcing through iterative design. The project produces new knowledge about crowdsourcing in the platform society, and has the potential to impact lives of crowdsourced workers around the world.