Workers in China worry over being replaced with AI: 'Everyone else is using it'

TEXT START: Computer programmer Fei Zhaojun's boss asked him if artificial intelligence could soon replace humans in coding jobs.

The Dissection

This article is an early autopsy of the wage-to-consumption circuit disguised as a labor-market feature. It records the sequence clearly: AI becomes adequate for mid-level cognitive tasks; firms cut staff; wages collapse; threatened training programs lose value; workers adopt AI because refusing is competitively suicidal. “Everyone else is using it” is the real thesis: adoption is not voluntary progress but an arms race forcing workers to help destroy their own bargaining power.

Its softening mechanism is individual adaptation—vlogging, training models, freelance studios, and using AI in teaching—plus a demographic escape hatch in which fewer workers supposedly make automation beneficial. Those are lag defenses and temporary niches, not restoration of mass productive participation. Human judgment survives only where it remains scarce, accountable, or physically embedded. Those are bottlenecks, not permanent moats.

The Core Fallacy

The article confuses higher productivity and lower labor demand with a healthier economy. Automation can offset a shrinking workforce in selected physical or care tasks while simultaneously eliminating the wages that sustain consumption. “AI as a tool” is not neutral under competitive ownership: it is a labor-saving capital asset. A worker may temporarily produce more with it, but the employer can then standardize the process, deskill the role, and remove the worker. Human oversight lasts only until fewer validators, automated verification, or liability procedures make it unnecessary.

Hidden Assumptions

• Displaced workers can continually find new niches at scale.
• AI errors guarantee humans remain indispensable rather than creating demand for fewer, cheaper validators.
• Demographic aging will absorb displaced labor instead of increasing the premium on automation.
• Productivity gains will flow into wages and broad consumer demand.
• Official unemployment captures wage cuts, unstable gigs, involuntary exit, and the disappearance of career ladders.
• State-backed AI diffusion can preserve social stability after productive participation erodes.
• National technological competitiveness translates into worker security.
• Independent ventures can compete with AI-equipped firms controlling capital, data, platforms, and distribution.

Social Function

Partial truth, transition management, and ideological anesthetic.

The article accurately documents early displacement and competitive compulsion, then domesticates the threat into reskilling, entrepreneurship, and demographic adjustment. It largely avoids the decisive question: who owns and controls the AI, and who receives the output after labor is no longer necessary? That omission converts a structural reallocation of power into a personal adaptation story.

The Verdict

This is evidence for the Discontinuity Thesis, not a refutation. China is already displaying P1 in exposed cognitive sectors, while state-backed diffusion weakens the possibility of preserving human-only economic domains. P3 is visible in layoffs, wage collapse, abandoned language programs, and fear-driven restraint in consumption, although the supplied text does not yet prove majority-level collapse.

An aging population may delay the break by sustaining demand for automation in labor-scarce physical and care work. It cannot repair the severed mass employment → wage → consumption circuit. “Use AI because everyone else does” is not a survival solution. It is the behavioral signature of a system entering self-liquidation.