'Street-­level Algorithm': Two Styles of Automated Law Enforcement by the Social Monitor Application

Abstract

During the COVID‑19 pandemic, many states actively began to create various apps to surveil citizens. One of the most famous such experiments in Russia was the 'Social Monitoring' (SM) application for monitoring the self-isolation regime for coronavirus patients in Moscow. Immediately after its appearance, the algorithm was heavily criticized, both for its technical quality and for the conceptual idea that the app independently decides whether or not to fine a patient for a violation. This design effectively meant that it became an automated substitute for a street-­level bureaucrat enforcing the rules. However, can we say that it, just like humans, can have its own style of law enforcement? In this article, we offer a conceptual description of the application as a sociotechnical ensemble of relations performing state executive functions – 'street-­level algorithm.' Using qualitative interviews with SM users, street-­level bureaucrats, and its creators, we illustrate two styles of SM enforcement inherent in the first and subsequent waves of the pandemic: hard-enforcement and soft-enforcement. Finally, we show that a hard-enforcement style with automatically imposed sanctions, in a bundle with a technically flawed appeal mechanism, is costly not only for users, but also for the authorities (overburdening state agencies and courts) and reduces the level of trust in a pandemic emergency. Soft- enforcement, coupled with spoken care, is a more comfortable control mechanism, especially in a situation of a greater flow of patients.