Curious about DataWorks? Here’s an article and a podcast about the organization. And the critical data literacy curriculum is available here.
Poster: “Calculating the cost of data refusal”
Abstract:
How does data refusal fit into the economic structure of a pro-social data work enterprise? If data is seen as “the new oil,” how can we argue for the value of refusal to engage in potentially harmful data work, in contestation of prevailing conceptions of data as a commodity? In this presentation, I will explore the implications of data workers’ refusal to continue working on a project that capitalized on their identity with concerning social implications. Reporting from DataWorks, a combined work training program and data services provider, I will share an ethnography of data refusal, stemming from a worker-centered training on critical data literacy.
Pulling from the larger theory of feminist refusal, I will employ the feminist data manifest-no to understand the tensions between economic productivity and socially just data creation and curation practices. Namely, as a small social enterprise with a limited budget, how do we, at DataWorks, weigh a given act of data refusal, and its implications, against a larger mission?
As a critical data scientist, my concern is both with the implications for data workers, as individuals, and the resulting dataset itself. In critical data studies, we know that pro-social and just treatment, including right to refusal, of data workers results in datasets that of higher quality (in terms of requester specifications), but this often runs counter to the expectations of ML and AI dataset requesters. In this presentation, I will share my work translating concerns of data workers as domain experts to machine learning practitioners.
Leave a Reply