With the growing adoption of cloud-fog computing, ensuring efficient task scheduling is essential to optimize energy consumption while managing workloads across distributed resources. Cloud computing provides high computational power but suffers from high energy consumption and network latency, whereas fog computing reduces response time by processing data closer to the source, but has limited resources. The integration of cloud and fog computing requires intelligent scheduling techniques to balance workloads efficiently while minimizing energy usage. This study presents a comparative analysis of four AI-driven scheduling models—CNN Scheduler (HunterPlus), Gated Graph Convolutional Network (GGCN), Bidirectional GGCN (Bi-GGCN), and Gradient-Based Optimization (GOBI). Each model employs a distinct AI-based approach to task scheduling in cloud-fog environments, focusing on optimizing resource allocation while reducing overall energy consumption. The results demonstrate that AI-based schedulers effectively improve energy efficiency, outperforming traditional scheduling techniques in dynamically changing workloads. However, scalability and adaptability remain key challenges when managing large-scale distributed systems.