彭博分析千万岗位 AI 暴露职位降幅高 10 个百分点
彭博对 2021 年 7 月至 2025 年 6 月近 1000 万条在线招聘岗位的分析显示,AI 暴露度高的职位数量降幅比低暴露职位多约 10 个百分点,入门级岗位与创意行业受冲击尤为明显。在线招聘总量在 ChatGPT 发布后的三年内减半,但分析人士提醒不宜过快归因于 AI,商业周期与远程办公等因素同样可能起作用。
报道正文
本文目录3 个章节
Bloomberg analysis of 10 million job postings shows AI-exposed roles in China fell 10 percentage points faster than others

An analysis of 10 million job postings by Bloomberg suggests that AI is putting additional pressure on China's already weak labor market. Entry-level workers and the creative industry in particular are reportedly already feeling the shift.
China's labor market has struggled for years with a sluggish economy, and AI now appears to be making things worse. According to a Bloomberg News analysis of nearly 10 million online job postings between July 2021 and June 2025, listings for jobs highly exposed to AI fell about 10 percentage points more than listings for less affected roles.
Bloomberg found that the total number of online job postings halved in the three years after June 2022, a starting point shortly before OpenAI launched ChatGPT and Chinese models like DeepSeek and ByteDance's Doubao rose to prominence. In nearly every industry, the decline hit jobs harder when their tasks overlapped heavily with AI capabilities, as measured by an index from Peking University. Bloomberg acknowledges, however, that the pure AI effect cannot be isolated with certainty because the broader economic slowdown may have hit office jobs disproportionately on its own.
Chinese studies back up the findings
Two independent Chinese studies reached similar conclusions, according to Bloomberg. Data scientist Chen Qin of ShanghaiTech University analyzed 116 million online job postings and estimates that AI caused a net loss of 5.2% in new hires between 2023 and early 2026.
A study involving the Chinese Academy of Social Sciences goes even further, finding 21% fewer online job postings and 4% lower average wages. Bloomberg Economics estimates that roughly 70 million workers in China are at risk from AI disruption, compared with about 45 million in the US.
Chinese tech companies avoid citing AI openly as a reason for layoffs because they fear backlash from the government and the public, Bloomberg reports. Companies including Xiaomi, Meituan, and Alibaba's travel unit Fliggy have denied mass layoffs. Employees at large firms, however, have reported waves of cuts since spring that affected up to 30% of some departments.
Previously separate roles are also merging, according to Bloomberg. One Beijing-based internet company reportedly combined programming, design, and testing so that a single person using AI would take on the work of up to six people. Job postings from Tencent for an "AI Design Engineer" and from BlueFocus for an "AI Director" point in the same direction.
AI micro-dramas are replacing film crews
According to Bloomberg, the shift is even more dramatic in Chinese micro-dramas, 90-second soap operas made for smartphones. Since the launch of new AI video generators like ByteDance's Seedance 2.0 in February, production time and costs have plummeted. Official data show that more than 90% of the 430,000 micro-dramas produced in the first eight months of 2026 were AI-generated. Linying Studio near Hangzhou, which opened only in late 2023, now sits mostly empty, and actresses like 26-year-old Zhang Ziqi report losing more than a third of their income.
According to the World Bank, China created only 21 million net new jobs in the five years through 2024, while youth unemployment reached nearly 19% over the summer. At the same time, more than 70 million college graduates are crowding into the job market, leading many to turn to retraining programs. China's leadership announced a nationwide survey on AI's labor-market effects in June, called for a "prudent pace," and overhauled about a third of degree programs. Former central bank adviser Cai Fang nonetheless warns that the current situation is only a "temporary lull" before a bigger AI shock.
Similar patterns among entry-level workers are emerging outside China, though under different economic conditions. Stanford University researchers analyzed millions of US payroll records and found that employment among 22- to 25-year-olds in highly AI-exposed occupations has declined 13% in relative terms since late 2022, while experienced workers held steady. The decline was limited to occupations where AI directly automates tasks, which the researchers attribute to models mainly replacing formal textbook knowledge rather than hands-on experience.
In Australia, too, an analysis by the central bank found weaker employment figures and higher joblessness among 15- to 24-year-olds in highly exposed fields compared with older workers. Like Bloomberg, however, the bank cautioned against drawing causal conclusions too quickly, citing possible business-cycle and remote-work effects.
{ "@context": "https://schema.org", "@type": "NewsArticle", "headline": "Bloomberg analysis of 10 million job postings shows AI-exposed roles in China fell 10 percentage points faster than others", "datePublished": "2026-11-20T16:56:03+01:00", "dateModified": "2026-11-20T16:56:14+01:00", "author": { "@type": "Person", "name": "Maximilian Schreiner" }, "publisher": { "@type": "Organization", "name": "The Decoder" }, "description": "A Bloomberg News analysis of 10 million job postings suggests AI is compounding China's already weak labor market. Entry-level workers and the creative industry are feeling the shift most, with more than 90% of Chinese micro-dramas now produced using AI. Former central bank adviser Cai Fang warns that a larger AI shock is still ahead.", "articleBody": "An analysis of 10 million job postings by Bloomberg suggests that AI is putting additional pressure on China's already weak labor market. Entry-level workers and the creative industry in particular are reportedly already feeling the shift.\r\n\r\nChina's labor market has struggled for years with a sluggish economy, and AI now appears to be making things worse. According to a Bloomberg News analysis of nearly 10 million online job postings between July 2021 and June 2025, listings for jobs highly exposed to AI fell about 10 percentage points more than listings for less affected roles.\r\n\r\nBloomberg found that the total number of online job postings halved in the three years after June 2022, a starting point shortly before OpenAI launched ChatGPT and Chinese models like DeepSeek and ByteDance's Doubao rose to prominence. In nearly every industry, the decline hit jobs harder when their tasks overlapped heavily with AI capabilities, as measured by an index from Peking University. Bloomberg acknowledges, however, that the pure AI effect cannot be isolated with certainty because the broader economic slowdown may have hit office jobs disproportionately on its own.\r\nChinese studies back up the findings\r\nTwo independent Chinese studies reached similar conclusions, according to Bloomberg. Data scientist Chen Qin of ShanghaiTech University analyzed 116 million online job postings and estimates that AI caused a net loss of 5.2% in new hires between 2023 and early 2026.\r\n\r\nA study involving the Chinese Academy of Social Sciences goes even further, finding 21% fewer online job postings and 4% lower average wages. Bloomberg Economics estimates that roughly 70 million workers in China are at risk from AI disruption, compared with about 45 million in the US.\r\n\r\nChinese tech companies avoid citing AI openly as a reason for layoffs because they fear backlash from the government and the public, Bloomberg reports. Companies including Xiaomi, Meituan, and Alibaba's travel unit Fliggy have denied mass layoffs. Employees at large firms, however, have reported waves of cuts since spring that affected up to 30% of some departments.\r\n\r\nPreviously separate roles are also merging, according to Bloomberg. One Beijing-based internet company reportedly combined programming, design, and testing so that a single person using AI would take on the work of up to six people. Job postings from Tencent for an \"AI Design Engineer\" and from BlueFocus for an \"AI Director\" point in the same direction.\r\nAI micro-dramas are replacing film crews\r\nAccording to Bloomberg, the shift is even more dramatic in Chinese micro-dramas, 90-second soap operas made for smartphones. Since the launch of new AI video generators like ByteDance's Seedance 2.0 in February, production time and costs have plummeted. Official data show that more than 90% of the 430,000 micro-dramas produced in the first eight months of 2026 were AI-generated. Linying Studio near Hangzhou, which opened only in late 2023, now sits mostly empty, and actresses like 26-year-old Zhang Ziqi report losing more than a third of their income.\r\n\r\nAccording to the World Bank, China created only 21 million net new jobs in the five years through 2024, while youth unemployment reached nearly 19% over the summer. At the same time, more than 70 million college graduates are crowding into the job market, leading many to turn to retraining programs. China's leadership announced a nationwide survey on AI's labor-market effects in June, called for a \"prudent pace,\" and overhauled about a third of degree programs. Former central bank adviser Cai Fang nonetheless warns that the current situation is only a \"temporary lull\" before a bigger AI shock.\r\n\r\nSimilar patterns among entry-level workers are emerging outside China, though under different economic conditions. Stanford University researchers analyzed millions of US payroll records and found that employment among 22- to 25-year-olds in highly AI-exposed occupations has declined 13% in relative terms since late 2022, while experienced workers held steady. The decline was limited to occupations where AI directly automates tasks, which the researchers attribute to models mainly replacing formal textbook knowledge rather than hands-on experience.\r\n\r\nIn Australia, too, an analysis by the central bank found weaker employment figures and higher joblessness among 15- to 24-year-olds in highly exposed fields compared with older workers. Like Bloomberg, however, the bank cautioned against drawing causal conclusions too quickly, citing possible business-cycle and remote-work effects.", "isAccessibleForFree": "False", "hasPart": { "@type": "WebPageElement", "isAccessibleForFree": "False", "cssSelector": ".paywall-active" } } THE DECODER+