On October 23, the General Office of the State Council issued its Opinions on Adapting to the Development of AI Technology and Building an Employment-Friendly Model of Development. The stated goal is blunt: strengthen the positive effects of AI in creating jobs and upgrading traditional positions, effectively ease the job-displacement effect of AI adoption, and keep technological progress, industrial upgrading, employment and people's livelihoods in step with one another.
Going by the contents compiled from public reports, the measures fall into five groups. The first three are about creating more jobs; the last two are about losing fewer of them and getting a clear picture of what is happening.
Where new jobs will come from
The first group is entrepreneurship. The document calls for the National Venture Capital Guidance Fund and the National AI Industry Investment Fund to play their part, backing corporate innovation through market-based means, and for a high-level open-source AI community, with a batch of open-source projects deployed to drive startups and jobs. It also names four models: "science and technology achievements plus entrepreneurship," "industrial development plus entrepreneurship," "vocational skills plus entrepreneurship" and "people's livelihood needs plus entrepreneurship," and asks for incubation services built around application scenarios and the commercialization of research results.
The second group is new occupations. The guidance asks for new occupations to be developed and occupational standards published, built around specific scenarios and large-scale deployment, to expand demand for roles such as digital management specialists, AI engineering and technical personnel, and AI trainers. It encourages local governments to add matching professional-title review tracks and to carry out skill-level certification promptly. These job titles have been entering the national occupational classification over the past few years, so the emphasis this time falls on "expanding demand" and professional titles. For people already doing data labeling or model tuning, a professional-title pathway is worth more than the job title itself.
The third group is traditional industries. Human-machine collaboration is to be promoted in the intelligent upgrading of industry and the digital and intelligent transformation of agriculture; producer services are to move toward greater specialization and the higher end of the value chain; and consumer services are to expand experience-based and personalized consumption, increasing the number of jobs the service sector can absorb.
How to soften displacement
The fourth group is the weightiest part of the document. It asks that AI be deployed first in roles facing labor shortages or dangerous environments, while the scope, intensity and pace of its use in labor-intensive industries are to be handled "prudently and steadily" so as to minimize the impact on workers.
The specific tool is the layoff buffer period: companies are encouraged to explore setting one up, give affected workers advance notice and provide support for retraining and job transfers. The wording is "encourage" and "explore," which makes these guiding provisions rather than requirements. How long a buffer should last, how wages are calculated during it and who pays for retraining are not spelled out in the public content; those details will have to wait for supporting rules from human resources authorities and local governments.
The fifth group is research. The document calls for special surveys that continuously track AI adoption, job creation and displacement, and shifts in skill demand; for a batch of AI employment research centers to be set up at universities and research institutions; and for projects funded by sources such as the National Social Science Fund to study AI's effects on employment and income distribution.
Who is most affected
Reading the fourth group alongside the third gives a rough sense of who will feel this most.
The first layer is employers in labor-intensive industries. The phrase about handling scope, intensity and pace "prudently and steadily" carries no binding force, but it signals a direction: replacing front-line jobs with AI on a large scale and all at once is likely to become politically sensitive. Outsourced customer service, telemarketing and basic quality inspection involve large headcounts and low barriers to replacement, so they are the likeliest to fit that language. This is our inference from the wording; the document itself does not name specific industries.
The second layer is people working in AI training, labeling and operations. Once occupational standards and title reviews for the new professions take effect locally, this group gets its first chance at a formal professional-title track, which would add a credential for job changes, residency applications and grading.
The third layer is research bodies and data. China currently has no public, regularly updated statistic on how many jobs AI has replaced. The document calls for special surveys and research centers; if the resulting data is published on a regular schedule, companies and job seekers would no longer have to judge which roles are shrinking from a feel for what they see on recruiting apps.
One caveat: the above draws on public reports that summarize the guidance, and the full text published on the State Council's website is authoritative. The document sets no target year and contains no specific quantitative indicators. Which provisions become enforceable rules depends on the detailed measures that departments issue next.
Sources: the Chinese central government portal (gov.cn), ITHome, CocoLoop; used to check the guidance's title, issuing body and the wording of the five groups of measures.