Benedict KingsburyRead PDFRead PDF
Digital engineers diagram ‘use cases’ to design software, based on practical needs of the quotidian product user rather than big normative claims. Human rights lawyers work in the reverse direction, starting from principles of universal application then applying these to hard cases. These two modes of thinking and practice have existed separately. Digital automation of government services using algorithms and AI is bringing them abruptly together and into mutual learning. This paper examines controversies and court decisions over digital welfare state programmes in Australia (Robodebt), the Netherlands (Syri), and the United Kingdom (Universal Credit), highlighted by Philip Alston as UN Special Rapporteur. The normative practice of human rights must grapple with data concentration and computerized decisions wherever power is exercised. The paper proposes ‘thinking infrastructurally’ as a path to bring human rights thinking into the fast-escaping public–private practices of algorithmic government and machine learning.
This paper was written as a chapter for The Struggle for Human Rights: Essays in Honour of Philip Alston (edited by Nehal Bhuta, Florian Hoffmann, Sarah Knuckey, Frédéric Mégret, and Margaret Satterthwaite, Oxford University Press.)