During the past decade, Artificial Intelligence (AI) has advanced rapidly, highlighting the need for AI education. To bridge the gap between theory and real-world application, we developed an interactive, web-based traffic management system that teaches AI-driven traffic management using computer vision and deep learning. Initially, a physical model simulating traffic scenarios was created, but limitations in video processing and hardware led to the design of a scalable web-based solution. This system detects vehicles and pedestrians under various lighting and weather conditions, such as fog, rain, and snow, offering an improved learning experience. Traffic management evolved from manual control to pre-programmed systems in the 1910s, which lacked adaptability. While modern AI-driven systems improve accuracy, they struggle in adverse conditions. Our system employs YOLO (You Only Look Once) object detection, prioritizing emergency vehicles while adapting to dynamic traffic scenarios. The system integrates a YOLOv8-trained detection module with transfer learning and data augmentation for robustness. A traffic management algorithm dynamically adjusts signals based on detected objects, allowing users to experiment with AI-based decision-making. A Python Streamlit-based web interface ensures ease of use in educational settings. This system enables students and researchers to explore AI applications in intelligent transportation by providing an interactive AI learning platform. It supports expansion to more road scenarios, allowing hands-on learning and testing before real-world deployment. Future enhancements aim to increase robustness and scenario diversity, strengthening its role as an AI education tool.