Artificial Intelligence (AI) in the form of different machine learning models is applied to Big Data as a way to turn data into valuable knowledge. The rhetoric is that ensuing predictions work well—with a high degree of autonomy and automation. I argue in this talk that we need to analyze the process of applying machine learning in depth and highlight at what point human knowledge production and negotiations takes place in seemingly autonomous work. Classification theory from Science and Technology Studies is able to make visible social dimensions of classificatory work and thereby highlight seemingly invisible knowledge production in the machine learning development and design processes. The talk suggests a framework for studying such classification closely tied to different steps in the work process and exemplify the framework on two experiments with machine learning applied to Facebook data from DATALAB – Center for Digital Social Research. By doing so the talk will demonstrate ways in which classification and potential discrimination take place in even seemingly autonomous models. Moving away from concepts of autonomy enable us to understand the underlying classificatory dispositifs in the work process and that this form of analysis constitutes a first step towards explainable AI and governance thereof.

Source: Bechmann, A. & Bowker, G. C. (2019) Unsupervised by any other name: Hidden layers of knowledge production in artificial intelligence on social media, Big Data & Society January-June: 1-11. https://journals.sagepub.com/doi/pdf/10.1177/2053951718819569

Anja Bechmann
Professor
School of Communication and Culture – Media Studies
Aarhus University