Abstract
Vector-borne diseases (VBDs), especially mosquito-borne infections, are shifting in range, seasonality, and intensity as climate variability, land-use change, urbanisation, mobility, population redistribution, and unequal control capacity reshape interactions among vectors, hosts, pathogens, and natural and social environments. Artificial intelligence (AI) is increasingly used for VBD surveillance, risk mapping, forecasting, and intervention planning, yet many applications are still judged mainly by predictive performance. The public health value of AI also depends on whether it improves decisions about when to act, where to intervene, which populations or ecological niches to prioritize, and how to allocate scarce resources under uncertainty. We synthesize the strengths, and limitations of AI applications across vector surveillance, risk assessment, early warning, and intervention decision support, with attention to dynamic at-risk population intelligence and local implementation. We argue that AI should be operationalized as part of a human-AI decision-support chain embedded in local disease prevention and control systems. Future efforts should strengthen multimodal data integration, dynamic population estimation, hybrid mechanistic-AI modelling, uncertainty quantification, implementation evaluation, and transparent governance. For VBD control, the priority is to embed interpretable, locally calibrated AI into accountable workflows that connect surveillance signals with timely, equitable, and feasible public health action.