Chill or Chilled? The Effects of Facial Recognition Technology on Democratic Participation and Social Discounting: A comparative analysis of two surveillance conditions

Abstract

The chilling effect, i.e. the indirect suppression of legally protected behaviour due to anticipated surveillance, has been extensively theorised but rarely empirically tested. Surveillance is largely treated as a uniform phenomenon. This study isolates two specific types of surveillance technologies — Facial Recognition Technology (FRT) and conventional video surveillance — to compare different surveillance modalities’ contribution to chilling effects. Drawing on chilling effect theory, the ontological distinction between representational and approximating surveillance, and social discounting research, we hypothesised that FRT would produce greater inhibition of democratic participation and steeper social discounting rates than video surveillance. In a between-subjects design experiment, participants (N=205) were exposed to either a (simulated) facial recognition surveillance condition or a video surveillance condition before completing a Participation–Self-censorship–Risk (PSR) survey and the Social Discounting Task–Short Form (SDT-SF). No significant differences were observed between conditions on either measure. These null results are interpreted and discussed in relation to Sweden's high institutional trust, cultural norms surrounding public accessibility of personal information, and methodological limitations of the experimental manipulation. This study contributes to human-centered AI and surveillance scholarship by empirically examining AI-driven surveillance while outlining extensive directions for future research on surveillance and chilling effects.