Into Hyperspace: A Critique of Hypernudge

Abstract

The term “hypernudge” has been offered by legal scholar Karen Yeung (2017) as a name for a growing phenomenon which sits at the intersection of behavioural science and computer science. Hypernudges are frequently described as personalised nudges, Big Data nudges, or data-driven nudges. Yet, at present, the concept of hypernudging has not robustly been integrated into the taxonomy of nudging described by nudge theory and behavioural science at-large. In this paper, I offer a critique of the hypernudging concept from a behavioural science and nudge theory perspective. The primary intention of this critique is to situate hypernudge as a distinct category within the existing nudge theory taxonomy, as well as to relate several ideas given by Yeung (2017) and associated authors to contemporary nudge theory. In doing so, I argue hypernudges produce, or at the very least magnify, three burdens which a decision-maker must overcome: the burden of avoidance, the burden of understanding, and the burden of experimentation. The burden of avoidance captures the costs facing a decision-maker who does not want to follow the hypernudge. The burden of understanding captures the costs facing a decision-maker who wishes to understand the dynamics of a hypernudge. Finally, the burden of experimentation captures the costs facing a decision-maker when merely interacting with a hypernudge, given the technical dynamics of the hypernudge. I argue that these three burdens can be understood through two concepts derived from nudge theory. The first of these concepts I call the assumption of error, whereby all decisions are assumed to be subject to some error which might be reduced through more nudging and more personalised nudging. The second of these concepts I call behavioural logic, which describes a system of logic – emergent from the assumption of error – which justifies various hypernudging practices. While I do not argue that either the assumption of error or behavioural logic are correct, I argue that insofar as a choice architect may believe them to be correct, they serve as a robust rationale for understanding the nature of hypernudges.