The UC Humanities Research Institute-funded Critical Machine Learning Studies working group is a collection of researchers across the University of California system.
This working group aims to articulate, from within the humanities, a technically specific and situated paradigm for engaging machine learning. Academic work on artificial intelligence seems now to be omnipresent, precisely because AI itself is such a rapidly developing field of research with widespread practical implementations lagging only slightly behind. The focus of many of these academic initiatives is genealogical, philosophical, and political. While we by no means eschew such perspectives, what we hope to contribute is a framework that yokes the theoretical and the practical by focusing on specific machine learning architectures, rather than understanding artificial intelligence as a technically and ideologically homogeneous cultural technique. We aim to leverage, in a collaborative and interdisciplinary way, the combined experience of UC faculty to design new approaches to describe, and critique in detail, contemporary machine learning systems, focusing on their individual (“ML studies”), rather than collective (“AI Studies”), properties, and their architecture-specific, rather than model-specific, biases and shortcomings.
An upcoming project is an addition to and intervention in the theorization of adversariality in machine learning systems.