ChatGPT has become popular around the world, with its strong contextual understanding ability and precise problem-solving skills, shocking the "people" as users. AI is becoming increasingly humanized, as if mastering mind reading skills and providing answers that are closer to human needs. The underlying logic of infinite proximity between humans and machines is that AI is infinitely approaching human thinking, and the mutual understanding between humans and machines is constantly improving. How do courtiers face the challenge of AI? How do we coexist harmoniously with AI? In fact, our exploration of coexisting with AI has precedent. Intelligent AI assistant judges and legal intelligent robots have greatly improved work efficiency in assisting trial activities and assisting people. The strong computing power and wide information tolerance of AI are breathtaking. The ultimate goal of AI is not to defeat people, but to assist them. In the complex trial work, AI has ample room for development: normalized and regular work is easy to be recognized and mastered by machines, and is more easily replaced by machines. For basic and repetitive work, we should be good at utilizing AI to improve efficiency. Utilize AI for basic data integration and assist in information correction. The formation of judicial documents requires a large amount of basic information integration, and AI can be fully utilized in information transformation to form a basic document framework; At the same time, utilizing the massive amount of documents and legal data stored in AI to provide intelligent case solutions for referee decision-making. How should judges coexist with AI? Hans Moravec once put forward the concept of topographic map of human ability, showing that some professions are on the edge of being submerged on the map. In the figure, from the bottom to the peak, different professions have different risks of being replaced, and professions at the bottom are most likely to be replaced by machines. Meanwhile, in the process of career substitution, new professions will emerge, and experts in each industry will not be replaced. On the other hand, in judicial work, both regularity and creativity coexist. For the parts with strong regularity, AI can master and provide auxiliary solutions through data source training; The parts with strong creativity still need to be integrated through the thinking of the human brain. In addition, the trial of a case is not only about drawing conclusions from documents, but also a process of determining and resolving disputes. In this process, more emotionally charged individuals have unparalleled advantages. At present, AI lags behind the human brain in creative work, and still relies on the support of judges in the combination of written legal provisions and vivid facts. Firstly, the overall control and decision-making of cases require legal thinking. The World Economic Forum has listed the ability to solve complex problems as the number one skill of humanity in the 21st century. Solving difficult problems is solving complex and uncertain conclusions. Any AI solution can only serve as a reference, and the independent thinking and decision-making of judges still dominate. There are similarities between cases, but they also have their own unique features. The answers provided by AI are limited to their understanding of the conditions, and the handling of the special conditions of the case is slightly insufficient. The judge's trial is not a mechanical task of drawing conclusions based on input conditions. It requires legal thinking to integrate and judge case information, as well as a comprehensive consideration of legal principles and human relationships. In this regard, AI has weak substitutability. Secondly, human emotions help resolve conflicts. Trial of cases