Safeguarding Data

Abstract

Abstract Over the past decade, data-driven systems have transformed how UK public services engage with the population. This ethnographic study, conducted in one of the United Kingdom's poorest boroughs, investigates the implementation of a machine learning data system for identifying at-risk children in need of safeguarding. It examines the concept of a ‘data consensus’ amongst council workers pursuing the ‘public good’. Within this consensus, the article explores contrasting stances on data's potential, namely that of data scientists, who use it as a tool to predict human behaviour and prevent harm, and social workers, who resist the notion that a person is predictable and therefore incapable of change. This reveals the political implications of these opaque algorithmic systems as various groups harness data in the name of public benefit.