Data forms of thought: “Unlearning Machine Learning”

I’m interested in showing both how power relations are intensified or reshaped by machine learning, and in how these techniques might induce us to think differently about forms of agency. Like many others, I’m increasingly aware of the need for critical work on the built-in differences and asymmetries of many data analytic and predictive systems.

My current book project into the data: unlearning machine learning considers some of the predictive products of machine learning from just this perspective. You can track the progress of the writing here data forms of thought. But it also attempts to find some ways in which these developments might be an occasion to rethink subjectivity, communication, media and power. That’s my slender hope.

The field of devices

I have a long standing interest in closing and software development as forms of work, labor, cultural practice and agency. I’m also keenly interested in how code and software transform worlds through redistributing and realigning relations and actions.

In recent work with Matthew Fuller and Andrew Goffey, as well as Richard Mills and Stuart Sharples, my attention head been focused on the site that hosts this blog: github. The metacommunities project studied the millions of code repositories in terms of how they affected each other through processes of imitation.

Data topographies and economics in genomics

I spent 2007-2012 in the Centre for Economic and Social Aspects of Genomics at Lancaster Cesagen. My research brief there was to study convergence cultures in genomics. This had two major effects on me. First, I encountered head-on the heavy traffic in biological data coming from DNA sequencing instruments moving towards databases. With Ruth McNally, part of my work in those years developed methods of making sense of large-scale data infrastructures both by working with practitioners and by treating the databases themselves as informants, as sources of statements or claims about data and its trajectories. Some of that work is beginning to appear in print.