AI in education is now everywhere in U.S. tech news. Even in China, I keep seeing stories about American professors realizing their students are using AI to cheat. At the time of writing, China’s largest exam, and for most teenagers its most important one, has been over for about a month.
My younger cousin has just survived the three-day ordeal known as the Gaokao, China’s national college entrance exam. He is now enjoying a rare two-month stretch of freedom: no homework, no teachers, no cram schools, no restrictions. Worried that he might have too little to do, my uncle has signed him up for driving school, hoping he can get his license in one shot.
But this is not really about my cousin. It is about AI and education.
According to China’s Ministry of Education, during the 14th Five-Year Plan period, from 2021 to 2025, Chinese universities added 10,200 new undergraduate majors while canceling or suspending 12,200 others. More than 30% of undergraduate programs were adjusted. In 2022, the ministry designated intelligent science and technology as a first-class interdisciplinary discipline, turning AI education into a national trend.
At the same time, China’s job market has become much harsher. The unemployment rate for people aged 16 to 24 stands at 15.6%. Chinese tech giants such as Xiaomi and Meituan are carrying out broad layoffs. Some of the pressure comes from AI itself: the boom has pushed up prices for upstream components, affecting the production of consumer electronics such as smartphones and reducing demand for labor. AI’s optimization of basic workflows also appears to be eliminating jobs companies now consider unnecessary. In China, the idea of a “stable job” has almost disappeared.
Universities are opening AI majors. AI, meanwhile, is destroying job opportunities.
The contradiction points to a larger failure. China’s long-standing model of using macroeconomic planning to forecast labor demand and guide talent training is being undermined by AI. The industry is changing so quickly that most university curricula are already detached from reality.
In a sense, AI is killing the Chinese university, or at least the version of the university Chinese families used to understand.
This article will explain that argument in three parts.
·Why choosing a university and a major in China is an almost irreversible investment.
·Why university majors in China function like investment products, and why AI is making those products unreliable.
·Why China may need a new index, or more precisely, a new kind of university.
Part One: In China, College Major Is a Stock You Can’t Sell
In China, getting into university and completing a four-year undergraduate degree is a high-cost investment with very little room for error.
Part of this mentality comes from the Gaokao. After the exam was restored in 1978, the number of students taking it rose year after year. This year, 12.9 million students took the test. They are ranked within their own provinces, which are roughly comparable to U.S. states, and then use their scores to apply for universities and majors.
The cutoff scores for Chinese university programs are not fixed in advance. They adjust based on the number and quality of applicants. Students submit their choices first, and only later find out the actual admission thresholds. Better universities, and majors tied to high-paying industries, usually require much higher scores.
To improve those scores, families invest heavily in tutoring, time, and every other available resource. If a student underperforms, the chance to enter a better major, attend a better university, and eventually move onto a better career track may disappear immediately.
Once students choose a major, transferring is rare. By the time they realize the curriculum or career path does not suit them, they may already be more than two years into a four-year degree.
Undergraduate education in China also does not work well as supplementary education for working adults. In theory, adults can retake the Gaokao and enter university later in life. In practice, doing so means paying additional tuition and housing costs while potentially going four years without stable employment.
Most professional credentials, including part-time graduate degrees, CPA certification, and CFA exams, assume the person already has a bachelor’s degree. Vocational colleges and junior colleges are still widely viewed as schools for students who did poorly.
In China, the university system resembles a stock exchange. Taking the Gaokao is like a company attempting an IPO, trying to list on the right board.
If your score is high, you enter a favored sector. You become part of the Nasdaq, or the S&P 500. More investors buy in. The company receives more capital. If your score falls short of the cutoff, you are out. It is basically an IPO failure — except for students, failing to get into a better university can be even more damaging than a company failing to go public.
That is why Chinese university majors are so tightly linked to better jobs. If a major no longer leads to stable or high-paying work, it quickly falls out of favor and is eventually cut. If it connects to a booming industry, a national strategy, or a hot labor market, it becomes a magnet for ambitious students.
A study by the education research firm MyCOS, based on suspension lists from 70 Chinese universities, found that marketing was the most frequently suspended major, with 16 universities halting enrollment. It was followed by public affairs management, logistics management, and IoT engineering.
The problem with many of these programs is that the skills they offer are too vague. If a company needs to study a market and develop a marketing plan, it does not necessarily need someone trained in a marketing major. Many management-related majors suffer from the same problem.
New industries, by contrast, often lack workers. That shortage gives companies room to offer higher salaries and more job openings. Haibao News, a Shandong-based news site, reported that AI workers in Beijing, Shanghai, Shenzhen, and Hangzhou earned an average monthly salary of 18,000 yuan in 2025. In second-tier cities such as Suzhou, the average was about 14,000 yuan.
To meet the needs of the new industry, new “artificial intelligence” majors are created. Changsha University of Science and Technology established a School of Artificial Intelligence last January by combining existing programs in automation, robotics, and AI. According to Fan Shaosheng, the school’s dean, graduates from its AI and automation programs have an employment rate as high as 90%, above the university average. China’s Ministry of Human Resources and Social Security has estimated that the country faces a shortage of more than 5 million AI professionals.
Part Two: AI Is Destroying the Investment Value of Degrees
Even without AI, China was already rotating through favored majors.
Thirty years ago, as China was preparing to join the World Trade Organization, international trade became a popular field. Twenty years ago, China’s internet boom made computer science highly desirable. Ten years ago, the real estate frenzy made civil engineering a coveted major.
The popularity of these majors was closely tied to macroeconomic policy. In the 1990s, China was opening up and absorbing global manufacturing capacity. In the early 2000s, it tried to follow the United States in building its own internet industry. Around a decade ago, in response to the global financial crisis, China launched a 4 trillion yuan stimulus plan that helped fuel the real estate boom.
But no major stays hot forever.
China’s real estate boom lasted until around 2016, yet civil engineering remained popular and required high entrance scores until about 2018, because real estate jobs were still seen as respectable and worth pursuing. In some provinces, students needed top-tier Gaokao scores to be admitted. Then the property market stalled. Suddenly, many of those students graduated into a much weaker labor market.
This reveals a crucial pattern: popular university majors usually follow the most profitable industries, and those industries are often shaped by cycles of government policy.
Within a typical five-year policy window, an industry receiving state support can become extremely profitable. One or two years later, universities begin launching majors tied to that industry. Three years after that, too many graduates enter the labor market and the field becomes saturated. Then the cycle begins again.
The enrollment trends of university majors often look like the price chart of a high-growth stock.
AI has broken this familiar cycle.
In theory, the first cycle of China’s AI undergraduate education should have ended around 2028. In 2018, the first 35 Chinese universities launched undergraduate AI majors. Since then, hundreds of universities, including vocational institutions, have rushed to open AI programs, even when they lack the faculty or infrastructure to support them.
Many universities, trying to attract new students and prevent older majors from being eliminated, have simply added the label “AI” to existing courses and repackaged them.
On Bilibili, the civil engineering program at Hefei University of Technology uploaded a promotional video for a program called “intelligent construction.” The video drew fierce criticism. Viewers accused the university of chasing trends and using a new name to disguise a major with weak employment prospects.
One commenter, who said they graduated in 2024, gave a detailed critique:
The major was still basically construction engineering, but with scattered electronics and computer courses added without foundations. Students had to study automatic control without first learning electrical engineering, write Keil and Proteus simulations without learning C, and take an “AI experience” course that consisted of labeling bounding boxes. “Isn’t that just treating students like temporary workers?” the commenter asked. “At least teach us Torch, Pandas, or TensorFlow.”
When a major marketed as “AI-related” does not teach basic mathematics, computer science, or programming, it may simply produce large numbers of unemployed graduates within three years. That will add even more pressure to youth employment.
But students who do study the foundations of AI face another problem: the commercialization of the industry is moving much more slowly than expected.
In just three years, OpenAI has gone from being seen as the global pioneer of AI to being accused of inflating an AI bubble. Anthropic, meanwhile, has helped fuel anxiety about AI replacing human workers, a fear that also brings the company media attention. Neither company has proved it can be sustainably profitable. Yet they have already pushed technology companies around the world to consider large-scale layoffs.
In the United States, Meta has chosen to cut around 8,000 jobs. In China, Meituan has been conducting mass layoffs since May while also deploying AI-model-driven customer service. Xiaomi has also carried out layoffs this year in its car and smartphone sales divisions.
At the same time, the companies producing this new technology are not absorbing young workers at meaningful scale. At least in China, DeepSeek only began publicly launching large-scale recruitment this year. It is unlikely to provide thousands of jobs.
The AI industry remains unstable. Unlike mature industries, AI has not yet produced a predictable labor market. That makes AI-themed university majors, hastily created to catch the wave, a poor guide to future employment.
We know that an export-oriented economy creates demand for international trade majors. We know that the rapid growth of the internet industry creates a large programmer workforce. Even a temporary real estate boom can create enough jobs for construction workers. But the application layer of AI is still changing at high speed.
No one is scheduling workflows around OpenClaw anymore, even though it was discussed only a few months ago. Cursor, one of the most popular AI coding tools in 2024, is already being squeezed by Claude Code.
As AI begins to replace many campus-recruiting jobs and internships, we still do not know which AI skills actually help people do real work.
The problem is that most AI-themed university programs can only teach students how to use AI tools. They do not necessarily produce concrete, job-ready skills that have truly been enhanced by AI.
Employers are responding with a new demand: they want candidates with real project experience who can also use AI quickly. In China, there is an old joke about companies looking for “fresh graduates with two years of work experience.” But the joke now reflects a structural problem.
As employers raise experience requirements, they are squeezing the job opportunities available to most graduates. The spread of AI is accelerating that process.
AI is now accelerating the collapse of a model Chinese families have long relied on: investing in knowledge to secure the future.
The “AI” major you study may have little to do with actual AI development. Even if you do study a serious curriculum, you may not find a corresponding job. Even if you find a job, you may not survive the next wave of layoffs in China’s emerging technology sector within three years.
It is like investors abandoning an index. If too many low-quality stocks are added to a prestigious benchmark, the index begins to decline. If the index keeps falling, investors eventually demand a new one.
Part Three: China Needs a New Index — or a New University
AI models themselves are already dissolving the existing form of university education.
For basic knowledge, AI can now provide abundant information while bypassing many traditional learning processes. From a teaching perspective, every major AI breakthrough forces textbooks to be rewritten. The university model people are familiar with is being eroded from both ends: research and instruction.
At this point, it is no longer very meaningful to ask whether AI has killed a specific major. The broader problem is that the standard university model — sitting in large lecture halls, listening to teachers explain prepared materials, taking regular exams, and eventually submitting a graduation thesis — is losing its purpose.
Without closed-book exams and mandatory assignments, almost every student can use Claude or DeepSeek to look up answers. In China, some students have even tried using large language models to write their graduation theses and then run them through plagiarism checks. That already violates the original purpose of university education.
I am not an education expert, so I went back to existing examples.
Wang Xingxing, the founder of Unitree Robotics, graduated from Zhejiang Sci-Tech University and later attended Shanghai University for graduate school. He was not a student at China’s most elite universities, but he was able to maintain an extraordinary level of focus on his robotics projects.
Lou Tiancheng, the co-founder of Pony.ai, came from the other end of the system. He was among the first graduates of Tsinghua University’s Yao Class, an elite program designed for top students in computer science.
That made me realize that the parts of university education most likely to survive may be the forms that involve real hands-on practice, such as building robots, or elite small-class programs for the very best students, such as the Yao Class.
Both models conflict with the current reality of Chinese higher education: mass enrollment expansion and dozens of students sitting in classrooms listening to lectures. They also suggest that many students may not need to enter a traditional university at all.
So what happens to everyone else?
If China had enough universities that offered credible vocational education, the employment pressure might be reduced. But in the Chinese public imagination, vocational schools are still where “bad students” go. Even when these schools can train highly skilled technicians, they remain overshadowed by China’s excessive pursuit of white-collar jobs.
China’s vocational education system is still underdeveloped. At the same time, comprehensive universities are losing value in both employment and academic research.
Some experts have argued that, for Chinese universities facing rapid technological change, regularly replacing old majors with new ones is only a stopgap measure. Chu Zhaohui, a senior researcher at the National Institute of Education Sciences, has noted that many of the majors being eliminated were only created in recent years and have had little time to mature. He argues that universities should give students more autonomy, allowing them to choose courses based on their interests, strengths, and career goals, rather than constantly reshuffling formal major structures.
I have not focused here on which humanities or arts programs are being cut, as if China simply does not value the humanities or the arts. But the pattern is visible there too.
Communication University of China, one of the country’s leading institutions for media and the arts, voluntarily canceled five majors, including photography, comics, and visual communication design. The university knows that photography or comics alone can no longer guarantee employment for its students.
Even among students in its broadcasting and hosting programs, many of whom once expected to become news anchors, the first job is often livestream e-commerce hosting. I know this because I have interviewed students like them.
There is a term in China for the degradation of today’s universities: “the high-schoolization of college.”
It describes the way universities are turning into intensified versions of high school, using textbook-based exams and internal grading systems to evaluate and manage adults as if they were still teenagers. In this environment, university students become more dependent on the school setting and more afraid of facing work and unemployment outside it.
Coincidentally, the rise of the term “high-schoolization of college” has occurred almost simultaneously with the mass adoption of large AI models in China. In a sense, it confirms how deeply AI is deconstructing the existing university system.
Conclusion
I have never attended an American university. But from available information, American universities are also trying to keep up with the AI boom.
The University of Pennsylvania’s engineering school began offering a Bachelor of Science in Engineering in Artificial Intelligence in the fall of 2024, calling it the first undergraduate AI degree in the Ivy League. Purdue University is going even broader: starting with the incoming class in fall 2026, all undergraduates on its main campus will have to satisfy a graduation requirement in “AI application capability.”
I recently had a video call with my cousin.
He has chosen education as his major and is planning to obtain a teaching certificate. In a country where the number of newborns is falling and the number of students is shrinking, becoming a teacher does not seem like the safest bet.
“But I think being a teacher is pretty good,” he told me. “At least you can pass something on.”
Incisive piece by X.Pin & CT Zhao on impact of AI on Chinese Universities.
Very nice piece! Education in every part of the world is being attacked by AI now. And no one has come up with a thoughtful plan to counter it.