The work aims to provide theoretical support for improving the generation of artificial intelligence-driven visual communication works, enhancing the quality of visual communication works and the efficiency of visual communication design. In this paper, relevant literature based on Scopus database and China Knowledge Network database were downloaded and collated to analyze the key technologies, research directions and research methods of existing artificial intelligence-driven visual communication work generation. The current research directions of AI-driven visual communication work generation are classified through intensive reading of literature, including text-to-image, image-to-image and video generation. The research methods adopted in each research direction are extracted, covering generative adversarial networks, knowledge inference, spatial adaption, etc. By analyzing the current status of research on AI technology-driven visual work generation as well as research directions, summarizing and summarizing the new step of research directions and methods, it opens up a new path for future designers to deal with complex design challenges, and at the same time provides references and bases for the future generation of AI technology-enabled visual communication works.