ABSTRACT Across 124 jurisdictions, formal legal architecture for AI‐related rights protection—including data protection legislation, independent oversight authorities, and sanctioning powers—is substantially more developed than the institutional conditions that make those rights operationally enforceable in practice. This paper introduces the concept of the capacity–justiciability gap to capture this structural disconnect: the distance between the formal existence of rights‐relevant legal infrastructure and the operational conditions necessary to activate it against algorithmic decisions and impacts. Drawing on an original dataset of 124 countries combining indicators of formal legal architecture with measures of operational institutional effectiveness, we document that this gap is widespread and cannot be explained by legal origin, political North–South classifications, or general measures of institutional quality. We further argue that two structural mechanisms—technological dependency and cost externalization—operate specifically on the operational side of the gap, eroding the effectiveness of oversight, enforcement, and redress mechanisms that formal frameworks establish but cannot guarantee in practice. These mechanisms resist quantitative operationalization with available comparative data and are examined qualitatively through the cases of Bangladesh, Brazil, and Kenya, selected as paradigmatic instances where substantial formal capacity coexists with limited operational justiciability. The findings suggest that AI governance must be evaluated not by the density of its formal frameworks but by the strength of the foundational institutional conditions that enable enforceable rights, with direct implications for how regulatory effectiveness is conceptualized in contexts of structural technological dependency. More broadly, the capacity–justiciability gap may characterize any regulatory domain in which formal authority is systematically decoupled from the material conditions necessary to exercise it—making AI an extreme but analytically generative instance of a wider structural condition in regulatory governance.
García-Llorente et al. (Mon,) studied this question.