CREATING AN AI AGENT TO INCREASE PUBLIC SAFETY: AN ACTOR-NETWORK THEORY STUDY

In this paper, we engage an important, but less explored research agenda on how artificial intelligence (AI) can help address societal challenges such as public safety. Our study focuses on creating an AI agent with a video monitoring system to enable real-time detection and prevention of drowning accidents on harbor fronts in Denmark. Despite being united by the goal of saving lives, the initiative encountered challenges, setbacks, and moments of renewed opportunity for progress in its translation from ideas and prototypes to an operational AI agent. To understand the complexity of this initiative, we draw from Actor-Network Theory and a 10-year case study to analyze the crucial role of an evolving network of human and non-human actors with heterogeneous interests. We unfold the translation process through three phases — network formation, network fragmentation, and network stabilization – in which ordering effects emerged, faded, and re-established as the initiative progressed. By implication, we advocate for a renewed focus on Actor-Network Theory in the digital era as a valuable framework for understanding AI’s role in addressing societal challenges, moving beyond the prevailing emphasis on business optimization and performance.