In mid March of this year, ChatGPT, a chatbot program, launched the latest version of GPT-4, which supports multiple input and output formats, giving it stronger professional learning abilities. As a chat program based on artificial intelligence (AI) technology, ChatGPT has attracted widespread attention from all sectors of society since its launch. Some even exclaim that AI will replace humans in dominating the battlefield. What changes will ChatGPT bring to the military field? Will AI surpass humans? With these questions in mind, this article starts with ChatGPT and outlines a general military application context for AI for readers. The potential military value of ChatGPT has attracted attention due to the introduction of new technology RLHF. The so-called RLHF refers to optimizing model algorithms through human feedback, so that the output of AI models is consistent with human common sense, cognition, and values. Simply put, ChatGPT is "more human like" compared to past AI models. This "image" is mainly reflected in natural language processing, namely semantic analysis and text generation. In terms of semantic analysis, almost any user's question can receive an effective response, unlike many times in the past where "the donkey's lips are not the horse's mouth"; In terms of text processing, the answer to any question appears logical, clear in meaning, and fluent in writing. It should be said that this is a major breakthrough in the field of natural language processing. This technology can clearly be applied in the military field. Usually, intelligence integration systems based on ChatGPT technology can target massive amounts of information on the Internet and serve as virtual assistants to assist analysts in data analysis, improving intelligence analysis efficiency and mining potential high-value intelligence. During wartime, the intelligence integration system based on ChatGPT technology can automatically integrate a large amount of battlefield intelligence into a comprehensive battlefield situation report, reducing the workload of intelligence personnel and improving their intelligence analysis and plan planning capabilities in fast-paced battlefields. ChatGPT can also be used to implement cognitive confrontation. In the era of information and intelligence, the degree of digitization in various countries is generally high, which means that the speed of information exchange, viewpoint dissemination, and emotional infection among the public is faster, which means that there is more space for cognitive attack and defense. ChatGPT's powerful natural language processing capability can be used to quickly analyze public opinion, extract valuable information, or make false statements to disturb public sentiment; It is also possible to use subtle and complex cognitive attack and defense tactics to induce, deceive, and even manipulate the cognition of the target country's people, in order to undermine their government image, change their public stance, and even divide society and subvert political power, achieving the goal of "defeating others without fighting". It is reported that the natural language processing technology used by ChatGPT is the key technology developed in the concept of joint global command and control of the US military. On July 1, 2020, the RAND Air Force project team released a report titled "Joint Global Command and Control in Modern Warfare - Analysis Framework for Identifying and Developing AI Applications". According to the report, AI technology can be divided into six categories. As one of them, natural language processing technology has