Abstract

AI systems help to efficiently allocate scarce public resources and are at the core of many smart city activities. Yet, the same systems may also result in unintended societal consequences, particularly by reinforcing social inequalities. CAIUS will identify and analyze such consequences. To this end, we develop an innovative methodology combining expertise from computer science and social science. Using agent-based models (ABM), we analyze the effects of AI-based decisions on societal macro variables of social inequality such as income disparity. The data input for these ABMs consists of both Open Government Data, digital traces, and own surveys. The goal is to train AI systems to account for their social consequences within specific fairness constraints; this synthesis of ABM and fair reinforcement learning lays the groundworks for what we call “Impact-aware AI” in urban contexts. With CAIUS, we accompany two smart city applications planned by partners in the Rhine-Neckar Metropolitan Region: dynamic pricing of parking space and traffic law enforcement via Internet-of-Things sensors. Our results contribute to research of human-AI interaction and will be condensed into general guidelines for decision-makers regarding the ethical implementation of AI-based decision-making systems in urban contexts.