Algorithmic reading as a precursor to policing action: what to do when routine surveillance of citizens yields too many irregularities

• In Event: POL Panel 13. Criminal justice data collection, data analysis and surveillance

Abstract

Surveillance tools are increasingly integrated in the digital toolbox of public administration: with access to multiple databases, citizens’ data can be elicited, drawn, scrutinized, aggregated, and cross-checked across time and multiple data nodes. Often, such routine algorithmic exercises that delve into citizens‘ data (such as on income, on migration and residency status, on taxation and welfare, or on social media activities) can produce multiple results of irregularities, thereby creating suspicion for possible illegal, illicit or fraudulent activities on a large number of citizens. This can bring forth severe real-life consequences: in the Netherlands, the Children Benefits scandal of 2021 created false allegations against thousands of residents, mostly those with a migrant background; the US no-fly included multiple innocent individuals, whose name was added due to flawed data handling; and the UK ‘hostile enviroment’ policy involved the erasure of migrants’ records due to alleged irregularities, leading to denial of rights and deportations. In other cases, participatory and lateral surveillance practices produce multiple allegations of illegal activities, prompting controversial and selective policing actions.

This paper explores the existing strategies employed by public servants in tackling the large number of irregularities arising from algorithmic reading of citizens’ data; and the type, scope and orientation of the policing action that follows. Drawing on existing literature, media articles and a limited number of semi-structured interviews with policymakers, it presents a set of practical and ethical considerations to guide civil servants when encountering large-scale data irregularities, and how these should (or should not) be acted on.

Author

• Lior Volinz, Vrije Universiteit Brussel