Automated welfare state

The automated welfare state is the use of algorithmic systems to determine eligibility, distribute benefits, and monitor compliance in public assistance programs. The term is central to Virginia Eubanks's analysis of how algorithmic governance reproduces historical patterns of social control in the administration of poverty. Rather than making welfare administration more efficient and objective, automated systems — from eligibility algorithms to fraud detection models to predictive risk scores in child protective services — concentrate discretionary power in technical systems that are opaque, unaccountable, and structurally biased against the populations they serve.

The automated welfare state is not a neutral modernization of public administration. It is a political technology that transforms the relationship between the state and the poor. The Victorian poorhouse was a physical institution; the digital poorhouse is an algorithmic one. Both systems sort populations into categories of deserving and undeserving, and both systems use the language of moral assessment to justify structural exclusion. The difference is that the digital poorhouse operates at scale, without human contact, and with the legitimating veneer of mathematical objectivity. Eubanks's work demonstrates that the automated welfare state is continuous with predictive policing and credit scoring in its use of algorithmic classification to govern marginalized populations.