The reduction in postings most exposed to AI

According to a Bloomberg News analysis cited by The Decoder, between July 2021 and June 2025, online job postings for roles highly exposed to artificial intelligence declined by about 10 percentage points more than those for less-exposed roles, across a sample of nearly 10 million postings. The decline reportedly affected almost all sectors, with a greater impact where tasks overlap more closely with the capabilities of AI systems, as measured by a Peking University index.

The total number of online postings reportedly halved in the three years after June 2022, a period preceding the arrival of ChatGPT and the rise of Chinese models such as DeepSeek and ByteDance's Doubao. Bloomberg acknowledges, however, that the effect attributable solely to AI cannot be isolated with certainty, because the broader economic slowdown may also have disproportionately penalized office jobs.

Two independent Chinese studies accompany the analysis. ShanghaiTech University data scientist Chen Qin, examining 116 million postings, estimates a net loss of 5.2% in new hires between 2023 and early 2026. A study involving the Chinese Academy of Social Sciences, by contrast, finds 21% fewer online postings and average salaries 4% lower. Bloomberg Economics estimates that about 70 million Chinese workers are at risk of AI disruption, versus about 45 million in the United States.

Corporate restructurings and micro-dramas

Chinese technology companies, Bloomberg writes, avoid openly citing AI as the reason for layoffs for fear of reactions from the government and public opinion. Xiaomi, Meituan and Fliggy, Alibaba's travel unit, have denied mass layoffs; however, employees at large companies have reported waves of cuts since the spring that in some departments affected up to 30% of staff. Some roles that were once distinct are also reportedly merging: an internet company in Beijing reportedly combined programming, design and testing, entrusting to a single person with AI the work that previously required up to six. Tencent postings for an "AI Design Engineer" and BlueFocus for an "AI Director" point in the same direction.

The change is most evident in micro-dramas, 90-second series designed for smartphones. After the launch of new video generators such as ByteDance's Seedance 2.0 in February, production times and costs reportedly plummeted: official data indicate that more than 90% of the 430,000 micro-dramas produced in the first eight months of 2026 were generated with AI. The Linying studio, near Hangzhou and opened only at the end of 2023, is reported to be largely empty, and actresses such as 26-year-old Zhang Ziqi report income losses exceeding one third.

The employment picture and international comparisons

According to the World Bank, Bloomberg reports, China created only 21 million net jobs in the five years through 2024, while youth unemployment reached nearly 19% during the summer. More than 70 million new graduates are reportedly pouring into the market, pushing many toward retraining programs. China's leadership announced in June a national survey on AI's effects on the labor market, called for a "prudent pace" and renewed about one third of degree programs. Former central bank adviser Cai Fang warns, however, that the current situation is only a "temporary pause" before a broader shock.

Outside China, similar dynamics are emerging for early-career workers. Researchers at Stanford, analyzing millions of U.S. payroll records, found that employment among 22- to 25-year-olds in occupations highly exposed to AI fell by 13% in relative terms since the end of 2022, while experienced workers remained stable; the decline reportedly applies only to occupations in which AI directly automates tasks, especially those based on theoretical knowledge rather than practical experience. In Australia, an analysis by the central bank reportedly found weaker employment and higher unemployment among 15- to 24-year-olds in exposed sectors, but, like Bloomberg, the institution cautions against drawing hasty causal conclusions because of the possible effects of the economic cycle and remote work.

Source and attribution: this original summary was prepared by AICronaca based on the report by The Decoder ↗. Statements are attributed to the source; any informational limitations: The causal effect of AI on the decline in hiring cannot be isolated with certainty; the estimates come from analyses and studies cited by Bloomberg and not from independent verification. Data on layoffs and restructurings are based on employee accounts, while the companies cited have denied mass layoffs. Further verification would be needed on the methodologies and coverage of the studies..