Recently, there were media reports that Fudan University scored 661 points in the admission line for undergraduate history programs in Guangdong this year, with the lowest ranking being 88th. Among the 11 majors for history enrollment, 8 majors include "humanities+artificial intelligence"; The minimum admission score for the Accounting and Intelligent Accounting Innovation Class at Jinan University is 603 points... This year, many universities have launched a series of "humanities+artificial intelligence" projects, which have been favored by candidates. The combination of humanities and artificial intelligence projects can be said to be a highlight of this year's college enrollment. Beijing Normal University launches the "Chinese Language and Literature+Artificial Intelligence" dual bachelor's degree program; The School of Design at Huazhong University of Science and Technology focuses on "AI+cultural innovation" and lays out areas such as "AI+creative entertainment". Compared to traditional liberal arts majors, these composite programs are more likely to stand out among numerous liberal arts enrollment programs. In the era of AI, the topic of "what is humanities" often enters the public eye, and many universities choose to reduce the number and scope of humanities enrollment; But other schools are actively striving to address the marginalization crisis of liberal arts by setting up interdisciplinary programs, new liberal arts, and other means. The latter, to some extent, is the "reverse force" of reality on the structural adjustment of liberal arts majors in higher education, but from the perspective of disciplinary development laws, this may be in line with the path of knowledge evolution. Looking back at the development process of the big language model, its early sprouts originated from the collision of traditional linguistics and computer technology. The traditional theories of linguistics have provided nourishment for the development of computing, and the assumption of "finite rules generating infinite sentences" laid the formal foundation for computational linguistics. It was not until later, the addition of statistical theory promoted several paradigm shifts, which gave birth to the large language models that people use today. On the contrary, current language models are not perfect, and their results often lack depth and interpretability, making it difficult to meet the professional accuracy requirements of social science research and strategic decision-making services. This also provides new directions for linguists to research innovation. The development history of the big language model shows the law of mutual integration and evolution between liberal arts majors and computational science. Only by deepening our understanding of this law can humanities education find the right direction for future development. After all, whether it is the humanities or humanities education that is in crisis, these are two separate propositions. People seem to be more concerned about the lack of integration between liberal arts education and employment needs, as well as the lack of integration between different majors. Traditional liberal arts education usually emphasizes the inheritance of knowledge within fixed disciplinary boundaries, and classic research methods such as text analysis, field investigation, and historical research are regarded as the standard. Theoretical teaching also has considerable "authority". The relatively fixed research paradigm and outdated theories have gradually led to the loss of vitality in the flourishing development of humanities. In contrast, science majors have already formed interdisciplinary fields dominated by cutting-edge technology applications, with physics, chemistry, biology, and even each other as methodologies and application scenarios. However, within the humanities, such interdisciplinary integration driven by technology is not common. Embracing AI may not only help break people's prejudice against the uselessness of humanities, but also inject new vitality into humanities education and professional development, helping students examine the value of knowledge from different perspectives, form different thinking logics, and lay the foundation for growing into versatile talents. However, although the direction of higher education adjustment is good, we should also be wary of "AI" becoming a new recruitment gimmick. 'New liberal arts' is not simply a combination of professional names, the key lies in how to truly integrate liberal arts with AI. How many AI applications are included in professional courses, what is the AI background of the teaching staff, what level of understanding of artificial intelligence technology can be achieved, and can the course setting ensure that students understand and comprehend thoroughly? All aspects need to withstand the careful scrutiny of students and various sectors of society. In recent years, some universities have offered courses on "Introduction to Artificial Intelligence", which have been criticized for problems such as "outdated PPT" and "excessive content". It can be seen that how to truly integrate artificial intelligence into undergraduate liberal arts education requires universities to carefully polish their teaching projects and truly achieve the full chain improvement from knowledge system to practical application. (New Society)