i
News
News · 2026-11-20

Bloomberg finds AI-exposed job postings fell faster in China

Dataist Dataist

Bloomberg’s analysis of nearly 10 million online job postings found that from July 2021 to June 2025, postings for jobs with tasks particularly exposed to AI declined about 10 percentage points faster than postings in less affected fields. Early-career workers and people in creative industries are feeling the changes most. The declines cannot be attributed to AI alone: China’s economic slowdown may also have hit office jobs disproportionately hard.

Cover: Bloomberg finds AI-exposed job postings fell faster in China

What the job-posting data shows

In the three years after June 2022, the total number of online job postings in China fell by half. That date came before OpenAI launched ChatGPT; Chinese models, including DeepSeek and ByteDance’s Doubao, became widely known later.

In almost every industry, job postings declined more sharply when the work involved tasks that substantially overlapped with AI capabilities. Bloomberg assessed the degree of overlap using an index from Peking University. But the analysis cannot separate AI’s impact from the effects of the broader economic downturn.

Two independent Chinese studies cited by Bloomberg reached similar conclusions. Data scientist Chen Qin of the University of Shanghai for Science and Technology analyzed 116 million online job postings. He estimated that from 2023 to early 2026, AI led to a net 5.2% decline in new hires.

A study involving the Chinese Academy of Social Sciences found 21% fewer online job postings and a 4% lower average salary. Bloomberg Economics estimates that AI-related changes could affect around 70 million workers in China and about 45 million in the US.

10 millionjob postings in Bloomberg analysis
5.2%decline in new hires
70 millionworkers at risk in China
45 millionworkers at risk in the US

How work is changing at companies

Chinese tech companies are reluctant to cite AI as a reason for layoffs. Bloomberg says they fear provoking public and government backlash. Xiaomi, Meituan and Fliggy, Alibaba’s travel division, denied carrying out mass layoffs. Still, employees at major companies reported waves of layoffs since spring, affecting up to 30% of workers in some departments.

The jobs themselves are changing, too. Bloomberg reported that one internet company in Beijing combined programming, design and testing: one person now uses AI to do work previously handled by as many as six employees. Tencent’s “AI design engineer” and BlueFocus’s “AI director” roles also reflect the merging of different tasks.

AI is changing microdrama production

The changes are especially visible in Chinese microdramas: melodramas made for smartphones, each about 90 seconds long. After a new generation of video generators, including ByteDance’s Seedance 2.0, came out in February, production times and costs fell sharply.

According to official data, AI created more than 90% of the 430,000 microdramas released in the first eight months of 2026. The Linying studio near Hangzhou, which opened only in late 2023, now sits nearly empty. Actress Zhang Ziqi, 26, said she had lost more than a third of her income.

430,000microdramas in the first eight months of 2026
more than 90%created by AI
more than a thirdZhang Ziqi’s lost income

Pressure on early-career workers

According to the World Bank, China added just 21 million net jobs in the five years through the end of 2024. Youth unemployment reached nearly 19% in the summer. More than 70 million university graduates are entering the job market at the same time, and many are turning to retraining programs.

In June, China’s leadership announced a nationwide study of AI’s impact on the labor market and called for its careful adoption. The country also updated about a third of university programs. Cai Fang, a former central bank adviser, described the current period as a temporary lull before a more serious shock from AI.

Similar shifts among early-career workers are visible in other countries, though their economic conditions differ. Stanford University researchers studied millions of US wage records and found that since late 2022, employment among 22- to 25-year-olds in occupations particularly exposed to AI had fallen by 13% in relative terms. Employment among more experienced workers remained stable.

The decline affected only occupations where AI directly automates tasks. The researchers link this to models replacing formal, textbook-based knowledge first, rather than practical experience.

An analysis by Australia’s central bank also found weaker employment outcomes and rising unemployment among 15- to 24-year-olds in AI-exposed fields, compared with older workers. Like Bloomberg, the bank cautioned against drawing quick conclusions about the causes: the economic cycle and remote work may have influenced the results.

Dataist’s take: The studies point to particular risks for early-career workers and occupations where much of the work can be described as a formal procedure. But the data does not allow AI’s impact to be confidently separated from the broader state of the economy, so the scale of its independent effect remains unclear.

Daily AI news

Every day we pick what actually matters in AI and explain it plainly — no hype, no filler. Subscribe if you want to follow where the industry is going.

Only what matters — every day

Follow on X