How AI Helps, Hurts China's Job Market: Tang Min - Forbes

TEXT START: Worldwide anxiety about AI’s impact on jobs isn’t going to fade anytime soon, particularly among young people looking for entry-level work.

The Dissection

The article is an admission of entry-level labor-market compression disguised as a manageable transition. Its most revealing fact is not the creation of AI trainers or a projected global net gain; it is that tasks once distributed across three junior hires can now be handled by one AI-proficient employee. That is the mechanism of displacement already operating.

The article then relocates the crisis from capital ownership to education policy. Universities, students, and instructors are told to adapt faster, acquire AI skills, and build industry partnerships. The underlying economic question—how many people remain necessary when the same output requires fewer workers—is never resolved. It is merely buried under retraining, collaboration, and aggregate job-count rhetoric.

The Core Fallacy

It confuses new tasks and aggregate vacancies with restored productive participation.

A projected net creation of jobs does not prove that displaced Chinese graduates will obtain them. The roles may differ in skill level, location, timing, pay, credential requirements, and scale. The article supplies its own counterevidence: entry-level white-collar work is being compressed, 22–25-year-olds are the most exposed cohort, and 75% of undergraduates attend institutions with weaker brand capital and fewer routes into high-end employment.

Under the Discontinuity Thesis, this is the early form of P1 and P3. AI does not need to produce spectacular mass layoffs to destroy the wage circuit. Suppressed hiring, fewer junior openings, and one worker absorbing the work of three are sufficient. “Task reshaping” is the polite label for reducing the number of humans required per unit of output.

The proposed policy response also assumes P2 can be defeated by institutional coordination. But universities revise curricula over two or three years while AI cycles move faster. Ceremonial partnerships, unequal corporate investment, and the Matthew effect described in the article show that coordination is already failing where it matters most.

Hidden Assumptions

• Every displaced worker can become productive in an emerging AI occupation through retraining.
• Five million workers with AI-enabled skills can be absorbed at comparable scale, speed, and income.
• New occupations will be accessible to ordinary-university graduates rather than concentrated among elite institutions and existing AI firms.
• Aggregate global job projections apply cleanly to China’s specific youth cohort.
• A “job opening” is equivalent to stable, economically meaningful participation.
• Education can outrun a technology whose iteration speed already exceeds curriculum cycles.
• Industry partnerships will deliver substantive training despite firms bearing immediate costs and receiving delayed returns.
• Policy can redistribute opportunity without confronting ownership and control of the AI capital producing the displacement.
• The absence of mass AI-only layoffs means the system is stable rather than entering through hiring freezes and cohort exclusion.

Social Function

Primarily transition management and ideological anesthetic, with a substantial element of elite self-exoneration.

The article acknowledges the wound so it can contain the diagnosis. AI is presented as both the cause of shrinking entry-level work and the source of enough new opportunities to justify continued deployment. Responsibility is shifted toward universities that adapt too slowly and students who lack “AI-collaboration competencies,” while the owners of the productivity gains remain structurally absent.

Its useful content is real: entry-level compression, institutional lag, unequal campuses, and rising youth exposure. Its ideological function is to convert a possible participation collapse into a skills-upgrade program. That preserves the legitimacy of the existing order while the employment ladder is being dismantled beneath the people waiting to climb it.

The Verdict

This is a partial-truth containment memo. It correctly records the first fracture—fewer junior workers needed for the same output—but mistakes retraining and new job titles for a replacement of mass productive participation. China has not disproved the Discontinuity Thesis; it is displaying its early phase: the ladder narrows at the bottom, elite AI positions multiply at the top, and policy is asked to manage the gap after the economic circuit has already begun to break.