Traffic congestion is a common global phenomenon that causes increased pollution, delays, and accidents. Conventional traffic light control systems, based on preconfigured timers, have proven ineffective as they cannot be adjusted according to dynamic and real-time traffic situations. This paper presents an artificial intelligence-powered smart traffic light control system, which, through computer vision and deep learning methodologies, strives to improve traffic flow efficiency. Installed on a Raspberry Pi platform, the system takes and interprets real-time traffic images and evaluates the number of vehicles heading towards every intersection. This report uncovers that the system adjusts the duration of traffic light signals in real time to optimize traffic flow, minimizing waiting times and boosting overall efficiency. The artificial intelligence model has been created to evaluate traffic conditions based on a predefined severity scale, enabling the system's capacity to disperse congestion, enhance traffic management, and reduce emissions. By prioritizing green traffic light signals for those routes with higher traffic volumes, the system aims to enable smoother, faster travel while also mitigating the environmental impact. This innovative strategy illustrates the power of artificial intelligence in revolutionizing urban transport networks, providing an efficient and sustainable means of addressing the increasing traffic demands of today's cities.