China AI Job Fears Grow as Youth Unemployment Climbs to 18.9%

Fear of AI layoffs has arrived in China

In the United States, a lot of the AI debate is about whether the technology could one day threaten humanity. In China the worry is closer to home and more concrete: people are afraid of losing their jobs, and some already have.12 Coverage from the past month, including AP reporting carried by Quartz and longer pieces in The Atlantic and The New York Times, describes the same thing. Chinese workers are adapting to AI as fast as the state is pushing it into the economy, and many of them doubt that adapting will keep them employed.21219

The main argument of this piece is that China has become the clearest test of what happens when a government orders rapid AI adoption into a weak labor market. Beijing's answer so far is to push automation hard while trying to keep firms from passing the cost on to workers through layoffs. That combination may hold for a while. It has not dealt with the underlying problem, which is that there are too few entry-level jobs for the number of young people looking for them.

The numbers behind the worry

The data looks bad. The National Bureau of Statistics reported that unemployment among urban 16-to-24-year-olds, excluding students, rose to 18.9% in August from 17.9% in July.20 That matches the August 2025 figure, and both are the highest readings since Beijing changed its method in 2023. Under the old method, the rate peaked at 21.3% that June.2013 A record 12.7 million university graduates entered the job market this summer.1920 One analyst quoted by The Guardian estimated that once earlier graduates, unemployed postgraduates and returnees from abroad are added, more than 15 million people may be looking for work.11 The overall urban jobless rate also rose slightly, to 5.3%.13

AI adoption has moved very quickly by any measure. According to IDC figures cited by the AP, 9.6% of Chinese industrial companies used AI models and agents in 2024, and 47.5% did a year later.12 Humanoid robots now sort parcels at postal facilities, and autonomous delivery robots are spreading through cities.12 Citi has estimated that about 70 million Chinese jobs, close to 10% of the total, face a high risk of displacement, with workers in their twenties most exposed.25

The people behind the statistics

The individual stories show how the change actually feels. Fei Zhaojun, a 40-year-old Beijing programmer, lost his job two weeks after his boss began asking whether AI could do the coding. He was one of about 160 people let go together, and he told the AP that mid-level coders are now replaceable in most cases.12 He has since left tech and makes short-form videos.3

Creative workers have been hit too. The Atlantic reported that Xu Peng, an actor working in short dramas, lost two expected roles when the studio switched to AI-generated video.2 A voice actor has sued a voice-generation platform he says used his voice without permission.2 Chinese social media users have started saying they fear being "distilled," meaning an employer uses a worker's chats, documents and emails to train an AI replacement.2

The youngest workers may be in the toughest position. The New York Times described one graduate who sent more than 2,800 applications in a year, got fewer than 20 interviews and no offers. Another paid for an AI course and still heard nothing from about 150 applications.19 A software engineering graduate told the South China Morning Post that AI can do some coding tasks faster and better than she can.20

Where the coverage agrees and where it splits

The outlets agree on the basic facts: adoption is fast, youth unemployment is high, and anxiety is spreading. They disagree on how much AI itself is to blame.

The Guardian, the Times and the SCMP all treat AI as an accelerant rather than the root cause. An Economist Intelligence Unit researcher said youth unemployment has been stuck since 2020. In that view, the problem started with China's move toward advanced manufacturing, which created a gap between graduates' skills and employers' needs, and AI has made that gap worse.11 A New Zealand academic told the Times much the same thing: AI arrived at a bad moment but did not cause the underlying problem.19 The SCMP pointed to weak domestic demand alongside AI.20

The International Labour Organization's research supports the more cautious view. Its study of 21 Chinese companies and 1,591 professionals found that firms mostly used AI to change how work is done, pairing it with human judgment rather than replacing staff outright.4 One insurer's 300 customer-service employees went from handling 6,000 issues a day to 15,000. The figures were self-reported, though, and most firms had no systematic way to measure wider effects.4 Respondents were split: 47% expected AI to create more jobs than it destroys, but 39% expected their own income to fall.4 A separate ILO estimate found that 7.5% of jobs held by East Asians aged 15 to 29 are in occupations highly exposed to AI.13

The harshest account came from an IANS-distributed report citing a Nepal-based outlet. It claimed the Communist Party has no retraining or safety-net policy and hides real unemployment figures.16 It also said that in some large firms, the share of AI-generated code jumped from 20–30% in 2025 to 90% by mid-2026.17 That claim should be treated with caution. It comes through a secondhand chain, and the statement that there is no policy response conflicts with a record of official measures documented by several other outlets.

One data point complicates the doom story. Recruitment platform Zhaopin reported that graduate-targeted AI engineer job postings grew 31.1% year on year, and robotics algorithm postings rose 57%.15 At the same time, youth unemployment stayed high. Taken together, that looks like a mismatch between skills and jobs, not a simple collapse in hiring.15 The new roles exist, but as one academic told the Times, they mostly go to the small number of candidates who already have AI skills.19

Beijing's answer: automate, but don't fire

The distinctive part of China's story is how the state is responding. Beijing wants AI everywhere: the AI Plus initiative targets 70% adoption in key sectors by 2027 and 90% by 2030.29 But officials have also become worried about job losses. Matt Sheehan of the Carnegie Endowment traces the change partly to public anger over Wuhan's robotaxis, which led officials to take displacement, and public fear of it, seriously.22

Since then, policy has shifted step by step. The AI Plus guidelines called for assessing the employment risks of AI and steering innovation toward sectors that create jobs.22 In January 2026, the labor ministry promised specific policies on AI's impact on employment.21 The 15th Five-Year Plan, released in March, explicitly named AI as a source of employment disruption.22 In July, the labor ministry and three other agencies ordered a system to assess AI's employment impact.25

Courts have been the most direct tool. A Beijing arbitration case found that adopting AI is a voluntary business decision, not the kind of unforeseeable change that justifies ending a contract.22 A Hangzhou court ruled for a worker whose job checking large language models was taken over by AI, and the ruling was designated a model case.2128 A Guangzhou court found that a graphic designer replaced by AI had been dismissed illegally. Its judge wrote that employers cannot push the risks of normal technological change onto employees.28 State news agency Xinhua warned that firms treating AI as a way to cut staff damage both their own competitiveness and employee trust.2

Why this approach may not be enough

These measures protect people who already have jobs. They do much less for people trying to get one. Reporting suggests firms are adjusting quietly: they cut contractors and graduate hiring and let headcount shrink through attrition rather than announcing layoffs.25 A court ruling against unlawful dismissal does nothing for a graduate who never gets an offer. Enforcement is another weak point. A Cornell labor scholar said the Chinese state lacks the capacity, and the interest, to rigorously enforce its own labor laws.30

Money is also limited. Bolder proposals include iFlytek founder Liu Qingfeng's call for AI-specific unemployment insurance and a union official's idea of capping automation within individual job categories.24 Analysts note, however, that local governments with depleted budgets have far less room to act than when Beijing spent ¥100 billion on displaced coal workers in 2016–2020.24 The 200-million-strong gig economy has long absorbed people who can't find other work. Delivery robots and robotaxis now threaten that too.1811

The political stakes are what make all this matter beyond economics. Carl Minzner of the Council on Foreign Relations noted the political risk for a nominally Communist party that is seen siding with business to replace workers with robots.2 George Washington University's Jeffrey Ding described Beijing as balancing productivity gains against regime stability. If AI requires large-scale displacement, he said, leaders might have to slow adoption.2 The likeliest outcome is that China keeps automating while managing the pace, protecting existing jobs at the cost of some efficiency. The young people now graduating into this market are the ones least protected by that approach.