The rapid proliferation of video content has elevated video semantic understanding to a pivotal research area in computer vision. Concurrently, advancements in natural language processing—particularly with large language models like GPT—have significantly enhanced chatbot systems. This paper provides a comprehensive review of the integration of video semantic understanding and chatbot systems. We explore core tasks and models in video understanding, including video classification, action recognition, and spatiotemporal modeling, and discuss how these are fused with chatbot technologies such as language models. Integration techniques like vision-language fusion, and video question answering (VideoQA) are examined in detail. We identify technical challenges such as multimodal data integration, real-time video processing constraints, and the complexity of emotion recognition. Computational limitations and user interaction issues are considered to provide a holistic view of the field. By analyzing the strengths and weaknesses of existing models and methods, we offer insights into future research directions, emphasizing the need for continuous innovation to overcome current challenges and fully realize the potential of intelligent human-machine interaction.