As the annual national college entrance examination approaches, the subsequent process of selecting universities and majors will become a new source of anxiety for many families. In recent years, there has been a growing preference among students and parents for science and engineering disciplines—particularly so-called "hard skills" fields like computer science, artificial intelligence (AI), electronic information, electrical engineering, automation, and materials science. Conversely, the popularity of traditionally high-scoring majors in economics, management, and finance has significantly waned, with admission scores for finance and economics universities noticeably cooling. Many higher education institutions are also adjusting their undergraduate program structures, reducing some traditional liberal arts and business offerings while increasing programs in science, engineering, agriculture, medicine, and interdisciplinary fields. Intuition might suggest that STEM (science, technology, engineering, and mathematics) fields have become more crucial in the AI era. AI itself is a product of STEM; its advancement relies on STEM; national competition and industrial security depend on STEM; and the upgrading of future industries also hinges on STEM. However, for the vast majority of people, a STEM education is not inherently more important than a liberal arts or business education. In the age of AI, the opposite may well be true.
The term "liberal arts and business" used here refers broadly to disciplines contrasted with STEM, encompassing both humanities like literature, history, and philosophy, and social sciences such as economics, management, law, political science, sociology, education, and psychology. Their common focus is not on how the natural world operates, but on how humans think, act, cooperate, and conflict, and how culture, organizations, and institutions shape our lives.
How STEM and Liberal Arts/Business Influence the World Differently
Setting aside abstract theory and looking at simple reality, among jobs requiring a university education or higher, those that frequently utilize specialized STEM knowledge are often limited to a few highly specialized technical or research roles. For most people, their work relies less on specialized technical knowledge and more on the understanding of human society and the real world provided by liberal arts and business disciplines—specifically, the comprehension and judgment of people, organizations, rules, and contexts.
Why is this the case? The reason lies in the different ways these two types of knowledge impact human society. When distinguishing between STEM and liberal arts/business, we often think of differences in research subjects and methodologies. However, a key distinction also lies in how they act upon the real world. Many achievements in STEM can be productized, mechanized, engineered, and offered as professional services, influencing the world without requiring understanding from the general public. One can use a smartphone without understanding semiconductors, fly in an airplane without knowing aerodynamics, use navigation and search engines without grasping algorithms, and take medication under a doctor's guidance without knowledge of pharmacology. STEM knowledge can be encapsulated within products, machines, materials, infrastructure, and professional services, allowing ordinary people to benefit without understanding the underlying principles. Over the past two centuries, the most direct and significant reasons for the substantial improvement in human living standards and social welfare have been the tremendous advances in science and technology. Yet, only a tiny minority directly participates in scientific research and development; the vast majority can enjoy the benefits of technological progress without needing to understand the science and technology itself.
Knowledge in the liberal arts and business fields operates differently. Insights from philosophy, history, economics, management, law, sociology, political science, and psychology typically cannot influence the world and improve lives by being turned into products. Instead, they must affect the world through human understanding, acceptance, and dissemination. If policymakers and those involved in drafting policies have no understanding of economics, it is difficult to produce sound economic policies. Economics must influence the world by shaping human thought and cognition. Similarly, other social sciences must be studied and understood before they can influence human behavior and decision-making.
Some might argue that STEM training is not merely about learning scientific, mathematical, and engineering knowledge; it also cultivates abstract thinking, modeling capabilities, logical reasoning, and causal analysis—all critically important skills in modern society. This is true, but these are not unique to STEM education; philosophy and the social sciences are equally capable of fostering these abilities. In short, STEM primarily influences the world through "objects" and "systems," while liberal arts and business do so more through "people" and "institutions." This difference is not one of degree but of mechanism. Of course, this distinction is not absolute—interdisciplinary fields like medicine, public health, nutrition, and environmental science share characteristics of both modes of influencing reality. However, this does not change a fundamental fact: for the daily work and social lives of the vast majority, what is more frequently operative is not how much scientific and technological knowledge they possess, but rather their understanding of people and society, and their comprehension and judgment of real-world contexts, social relationships, and institutional environments.
The Proportion of STEM in China is Already Quite High
Statistics from China's Ministry of Education show that STEM students account for 40% of ordinary undergraduate graduates in China, a figure significantly higher than in many other countries. According to international data from the OECD, the proportion of STEM graduates from highest to lowest is as follows: Germany 35%, South Korea 32%, Canada 26%, Mexico 26%, United Kingdom 25%, Switzerland 24%, United States 24%, Italy 21%, France 20%, Japan 18%, Brazil 15%.
A common argument for continuing to expand STEM education is that China is a major manufacturing nation, and its "engineer dividend" is a competitive advantage, so naturally, STEM should have a larger share. Taking 2023 as an example (the latest year with comparable data), the percentage of manufacturing value-added to GDP is: China 25%, South Korea 25%, Japan 21%, Mexico 20%, Germany 19%, Switzerland 18%, Italy 15%, Brazil 13%, France 10%, United States 10%. From an international comparison, the relationship between manufacturing share and STEM graduate share is not straightforward. Countries with similar manufacturing shares, such as China and South Korea (both 25%), and Germany and Japan (19% and 21% respectively), still show large disparities in STEM graduate proportions: China's 40% versus South Korea's 32%, Germany's 35% versus Japan's 20%. The United States, with a relatively low manufacturing share (10%), still has a not-insignificant proportion of STEM students (24%), exceeding that of Japan, which has a much higher manufacturing share.
As a major manufacturing power, China certainly needs to value STEM education. However, compared with other nations, including those where manufacturing remains crucial, the proportion of STEM in Chinese higher education is already very high. In this context, whether to further compress liberal arts and business programs and channel more resources and attention towards STEM is a question worthy of deep reflection. The continued development of China's economy and the sustained upgrading of its industries depend not only on research personnel and engineering technicians but also on various types of managers and non-technical staff. It relies not only on manufacturing but also on the vastly larger employment sectors within various service industries.
China indeed needs STEM to tackle cutting-edge hard technology challenges. However, key technological breakthroughs primarily depend on high-level research talent; this does not equate to "the larger the scale of undergraduate STEM programs, the better." On the contrary, how to transform advanced technology into profitable products and industries is equally important and requires a large number of excellent management and business professionals.
AI is Reshaping the Value Hierarchy of Skills
The development of AI will lead to more and more technical capabilities being encapsulated into products and tools. Many tasks that previously required substantial professional training will increasingly be accomplished with the aid of AI, no longer remaining specialized skills exclusive to a small number of technical experts. The role of humans is unlikely to diminish, but the way humans contribute will change. As the ability to "get answers" becomes increasingly cheap and ubiquitous, what truly becomes more valuable are the abilities to define problems, set objectives, screen criteria, assess risks, and assume responsibility.
AI can help people search for information, organize materials, compare options, and generate reports more quickly, but it cannot decide for humans which questions are worth asking, which goals are worth pursuing, which costs are unacceptable, or which consequences must be guarded against. AI can participate in analysis, prediction, and suggestion, but it is not the entity that bears political, legal, organizational, or moral responsibility. Whether at the national level for economic and social policies, or at the corporate level for strategy, organization, and investment decisions, AI will become an increasingly important辅助工具 (auxiliary tool), but the final decisions and the consequences thereof will still rest with humans. As long as decision-making authority and responsibility remain in human hands, the decision-maker's own understanding of the decision's objectives, constraints, effects, and risks is indispensable.
Therefore, the AI era has made certain abilities more important than before. In the pre-AI era, competitiveness in many positions stemmed mainly from information processing, data collection, standardized expression, and basic analytical skills. As AI becomes increasingly proficient at providing these capabilities, what will truly differentiate individuals is their understanding of specific contexts, interest relationships, and social consequences, as well as their ability to make judgments and trade-offs under complex constraints. This is precisely where the value of liberal arts and business education, especially social science training, lies.
Economics helps people understand incentives, constraints, and trade-offs. Management helps people understand organization, coordination, and execution. Law helps people understand rules, boundaries, and responsibility. Sociology and political science help people understand institutions, power, and group behavior. History helps people understand path dependence and complex consequences. Psychology helps people understand cognitive biases, motivation, emotion, and judgment errors in uncertain environments.
In the AI era, the training within liberal arts and business that helps people understand goals, institutions, organizations, incentives, trade-offs, risks, and responsibilities has become more important than ever. If current liberal arts and business programs are not adequately providing this training, it precisely indicates that these disciplines themselves have significant room for improvement and require more investment in teaching and research resources.
The research, development, diffusion, application, and commercialization of science and technology have never been purely technical issues. What kind of institutions incentivize innovation? What kind of organizations can transform technology into products? What kind of market and legal environment allows innovation to diffuse? What kind of cultural and social structures influence people's acceptance or resistance to new technologies? These are not questions that can be answered solely through STEM training. The more technology needs to move out of the laboratory and into markets, organizations, and social life, the more unavoidable these questions become.
Therefore, even students in STEM majors should take some courses in liberal arts and business. Similarly, students in liberal arts and business should also take some STEM courses. In the AI era, all students should possess basic literacy in the humanities and social sciences, as well as basic literacy in science and technology. All students should learn some economics, sociology, law, psychology, history, and philosophy, and they should also learn some computer science, artificial intelligence, and data processing.
STEM education will, of course, always be important. However, as more and more technical capabilities are encapsulated into AI-enabled products, systems, and tools, the abilities to ask questions, set goals, understand constraints, weigh consequences, make judgments, and assume responsibility will become even more valuable. These are precisely the abilities that liberal arts and business education can help cultivate. It is in this sense that we argue the AI era may have an even greater need for liberal arts and business.
Comments