China has formally expanded its job catalog to reflect the growing presence of embodied AI and software agents in the workplace. On September 9, labor authorities issued an update that adds 11 occupations and 23 specialties to China’s existing occupational classifications. The list includes an “embodied intelligence robot technician” as a new occupation and an “AI agent developer” as a new specialty, according to the State Council’s English portal.
The difference between those labels matters. A new occupation is not the same as a specialty added beneath an existing occupation. China’s announcement recognizes embodied-robot work as a distinct job category while treating AI-agent development as a specialized capability within the broader labor classification system.
The release gives policymakers, vocational schools, companies, and workers a shared vocabulary for jobs that have emerged faster than many formal training programs. It does not mean that every company now has a standardized role or that new workers will immediately receive the same credentials. Authorities said they will develop national standards for vocational education, skills training, and talent assessment.
Eleven New Occupations and 23 New Specialties
The September 9 release adds 11 new occupations and 23 new specialties. China Daily reported that five of the new occupations are digital and three are green. The update brings the country’s total number of digital occupations to 113 and green occupations to 142.
The announcement is the eighth batch of new occupations released since 2019. With this batch, China has identified 121 new occupations in that period. Such catalog updates are a policy tool: they make it easier to define training, certification, education, and labor-market statistics around work that did not fit neatly into older classifications.
The embodied intelligence robot technician is the most direct recognition of physical AI in the release. The title suggests roles related to installation, operation, maintenance, testing, or servicing of robots that use perception and AI-based control. The official announcement did not publish a complete task list for the new occupation, so it would be premature to assume exactly which duties or qualifications every employer will require.
The AI agent developer appears in the list of new specialties. That reflects a different labor need: building, configuring, integrating, and supervising agent systems that use models, tools, and workflows. It also aligns with a market in which companies are asking for people who can turn language models into task-performing software.
EastFrontier earlier reported a surge in AI-agent job postings in China. The official classification does not validate every market forecast, but it does show that national labor policy now treats agent-related skills as concrete workforce needs.
Training Standards Become the Next Policy Step
The government said authorities will draft national standards for vocational education, skills training, and talent assessment. That next stage is more significant than the name list alone. A job title becomes useful to employers and workers when it is connected to curricula, equipment, examinations, credentials, and employment pathways.
For embodied intelligence robot technicians, training could eventually include mechanical systems, sensors, safety procedures, control interfaces, data collection, and troubleshooting. For AI agent developers, coursework may need to cover model use, tool integration, workflow design, data security, and evaluation. The specific standards have not yet been published, so these are practical areas the programs may need to address rather than confirmed curriculum requirements.
China has been building AI into education and vocational planning for several years. EastFrontier covered the policy push to integrate AI into national education reform. The September 9 announcement adds an employment-side counterpart: instead of only teaching AI concepts, the system is identifying job functions that need recognized skills.
The classification may also help companies communicate what they need. A robot manufacturer, warehouse operator, or service provider can refer to a recognized technician category. A software firm can identify agent development as a specialized skill rather than a vague request for “AI talent.” That clarity can matter for hiring, wage surveys, and training partnerships.
Formal Recognition Does Not Settle the Labor Impact
The new roles arrive as employers consider how AI changes existing jobs. A robot technician may maintain machines that reduce repetitive manual tasks. An AI agent developer may automate portions of office work. Formal recognition can create opportunities for workers to move into higher-skill roles, but it does not guarantee that all displaced or affected employees will be able to make that transition.
The policy’s value will depend on implementation. Schools and training providers need instructors, equipment, and up-to-date materials. Companies need to provide real jobs with career paths rather than use new titles as a branding device. Workers need access to training that fits their backgrounds and locations.
The announcement also has a measurement function. By defining occupations and specialties, authorities can gather clearer information about demand. That can influence where vocational programs are built, how local governments support training, and how companies report their employment needs.
China’s new catalog makes embodied AI and AI agents visible in the labor system. It does not determine how fast robots or agents will spread, nor does it resolve questions about wages, displacement, or workplace protections. It does establish that the skills needed to build, operate, and supervise these systems are no longer treated as informal add-ons. They now have a place in official workforce policy.