This paper presents a novel approach to improving urban road traffic control using artificial intelligence (AI) for dynamic traffic light management. We begin by describing the current context of urban traffic management and the challenges facing traffic light infrastructures. We then explain how AI can be integrated into this context for more effective regulation. We have chosen to use a basic model based on convolutional neural networks (CNN) to model road traffic in real time. This model collects real-time data from traffic cameras and other sensors, pre-processes it and then analyses it to make intelligent decisions about traffic light control. By using historical data and adapting to changing conditions, our model has been able to reduce waiting times at intersections, minimise congestion and improve traffic flow. This research paves the way for more intelligent and adaptive traffic management in urban environments. The practical implications of our approach include more efficient urban mobility, reduced greenhouse gas emissions and improved road safety. Future prospects lie in the continued optimisation of the AI model and its integration with other intelligent transport systems, contributing to more sustainable and liveable cities.