The Artificially Unintelligent Data Double

Abstract

This essay engages with Haggerty and Ericson's concept of the "data double" to examine how contemporary surveillance practices have produced increasingly powerful yet fundamentally flawed representations of individuals. While the surveillant assemblage remains central to modern life, today's data doubles are characterized by a paradox: they wield immense decision-making power in determining access to housing, employment, credit, and healthcare, yet are too often built on inaccurate, incomplete, and inaccessible information. Three U.S. case studies frame the contemporary data double as financial risk through big data credit scoring, as housing and employment applicant, and as vulnerable subject in health and welfare systems. These cases demonstrate how prioritization of data volume over precision has created "artificially unintelligent data doubles"—a systematically distorted abstraction that serves institutional profit and control rather than accurate representation. Opacity in these systems makes challenging one’s data double practically impossible, creating profound stratification as individuals face two untenable options: accept inaccurate representations that deny them opportunities, or actively construct manipulated data doubles designed to pass algorithmic screenings.

Keywords:

• data double,
• artificial intelligence,
• Surveillant Assemblage,
• opacity