Title: Integrating low-altitude technologies into safety governance in high-risk sectors: A STAMP-informed socio-technical framework and comparative case analysis
ABSTRACT: Low-altitude technologies are increasingly used to support safety inspection and risk monitoring in high-risk sectors. However, their safety contribution depends not only on technical performance but also on how technological information is incorporated into organisational and regulatory control processes within a hierarchical socio-technical system in which safety outcomes emerge through dynamic cross-level interactions. This study develops a socio-technical framework informed by the Systems-Theoretic Accident Model and Processes (STAMP) for analysing the integration of low-altitude technologies into safety governance. The framework comprises four interconnected levels: public governance and policy, the low-altitude technology ecosystem and infrastructure, enterprise safety management systems, and physical operations and operating environments. It examines safety constraints, downward control actions, upward feedback, controller responsibilities, and potential control weaknesses across these levels. The framework is applied to a theory-guided comparison of two contrasting cases in Guangdong Province, China: government-led unmanned aerial vehicle (UAV) inspection of work-at-height activities in Jiangmen and artificial intelligence (AI)-enabled UAV inspection of construction sites in Futian District, Shenzhen. The comparison identifies two integration pathways: inspection-based regulatory correction and platform-enabled dynamic intervention. The findings suggest that technological detection can support effective safety governance when responsibility for interpreting warnings is assigned to accountable actors, appropriate intervention follows, and corrective actions are carried through to verified closure.
Keywords: Low-altitude technologies;Safety governance; Systems-theoretic accident model and processes; Safety risk; High-risk sectors
DOI:https://doi.org/10.1016/j.ssci.2026.107534
