AI Surveillance Technologies in Smart Cities: Privacy Calculus versus Privacy Paradox

Abstract

Smart cities represent the convergence of technological innovation and urban development, aiming to enhance efficiency, safety, environmental sustainability, and overall well-being through interconnected systems, sensors, and digital devices. At the heart of these innovations lies the deployment of AI-powered surveillance technologies, which contribute to monitoring and managing urban environments more effectively. While such systems promise improvements in security and operational efficiency, they also raise pressing concerns about individual privacy and data security. This study examines the tension between technological progress and privacy preservation in AI-based surveillance systems, focusing on how citizens from seven smart cities perceive and respond to these technologies. Drawing on a quantitative pilot survey conducted in seven smart cities and using a five-dimensional framework of privacy concerns, this paper maps citizen attitudes towards AI surveillance technologies. These are cross-analysed against seven distinct categories of AI surveillance technologies deployed in public spaces. A central question of the analysis is whether individuals’ responses reflect the privacy calculus – a rational evaluation of risks and benefits, or are more consistent with the privacy paradox, where expressed concerns do not translate into protective behaviours, often due to insufficient awareness or a lack of options to opt out. In addition to assessing the overall levels of privacy concern, the study ranks five privacy dimensions based on the degree of concern they elicit and evaluates which types of AI surveillance technologies are most and least acceptable when privacy is factored into the adoption equation. We further introduce a privacy-weighted adoption attractiveness metric to measure public receptivity of the seven types of AI surveillance technologies. The findings, derived through descriptive statistical methods, reveal trends and peculiarities across the cities and the respondents’ demographic characteristics, such as gender, education level, and age. These insights contribute to a more nuanced understanding of how privacy values interact with the promises of AI surveillance in smart cities.