Abstract
Abstract
Artificial intelligence (AI) is rapidly becoming entrenched across government operations, transforming administrative processes and service delivery, particularly in smart cities and digitally advanced nations. While AI applications such as chatbots, predictive systems, AI of Things (AIoT) devices, robotics, and agentic AI systems promise greater efficiency, innovation, and decision-making support, they also raise salient questions about how these systems are governed and, in turn, how they shape governance. This study conceptualizes AI governance and integrates cross-sector governance risks and challenges to develop an adaptive, ethical and responsible AI governance framework. It defines AI governance as an institutional system that brings together legal rules, ethical oversight, organizational capacity, and human–AI decision-making arrangements to ensure the responsible use of AI that serves public values and attenuates systemic risks. The paper further elucidates distinctions between the “
governance
of
AI
,” which concerns the regulation and oversight of AI systems, and “
governance
by
AI
,” in which AI systems increasingly provide inputs into decisions that affect public life. The analysis identifies recurrent and pervasive cross-sector AI governance risks and challenges, including algorithmic bias, weak oversight, regulatory gaps and capture, vendor dependence, geopolitical and socio-political risks, environmental challenges, and legitimacy and trust deficits. Because governance arrangements themselves may produce new risks or unintended consequences over time, AI governance must remain adaptive and responsible rather than fixed. Building on digital era governance, adaptive governance, and polycentric governance, the paper proposes a six-step adaptive, ethical and responsible AI governance framework that embeds risk management, responsiveness, and ethics-by-design across the AI lifecycle. The framework is illustratively applied to New York City and Singapore to show how its six governance dimensions can organize analysis across a decentralized smart city and a centralized smart nation. Furthermore, it identifies transparency, accountability, public trust, human–AI collaboration, public value, and vendor oversight as core governance design principles for smart cities and nations adopting AI.