AI now makes 95% of China's short video series. Hundreds of thousands of jobs in the ...

TEXT START: The rapid adoption of artificial intelligence (AI) tools in video production and e-commerce in China is replacing human actors and presenters in a sector that directly employs hundreds of thousands of people.

The Dissection

This is a live substitution event, not a forecast. The text describes AI producing the overwhelming majority of short series, reducing production costs to roughly one-tenth and compressing a five-person, three-month process into one person working one or two days. Output expands while labor requirements collapse.

The e-commerce examples show the same mechanism applied to attention and sales: always-on synthetic presenters operate at a fraction of human cost, while digital replicas turn identity, voice, and performance into duplicable inputs. This is a direct P1 signal, with P3 already beginning in the affected labor pools.

The court case and image-rights disputes are P2 lag defenses. They may impose compensation or friction. They do not restore the workers’ economic necessity.

The Core Fallacy

The article treats displacement as jobs being “at risk” while treating increased output and productivity as inherently beneficial. It never follows the decisive chain: cheaper synthetic labor reduces the wage income of human workers, which weakens the consumption circuit that post-WWII capitalism depends on.

The claim that 95% of productions use AI does not prove that 95% of jobs have vanished; supervision, editing, taste, rights clearance, and platform control may remain. But those residual functions require far fewer people and increasingly concentrate power in owners and controllers of AI capital. More content does not mean more mass employment.

Hidden Assumptions

• Expanding content demand will absorb displaced workers. Supply can grow faster than attention, pushing prices and labor income downward.
• The remaining human tasks will be numerous and well paid. In practice, they become concentrated controller or servitor roles.
• Legal compensation will preserve human careers. It can transfer rents or delay deployment, but cannot make human labor cost-competitive again.
• State-backed productivity will translate into broad prosperity. Beijing’s support accelerates the automation mechanism; it does not guarantee distribution of its gains.
• Human authenticity and trust form a permanent moat. They may survive as premium niches, but platform economics will continuously synthesize and cheapen them.

Social Function

Classification: partial truth functioning as transition management and ideological anesthetic.

The article supplies useful evidence of replacement, but its language—“at risk,” “productivity,” and “new sources of economic growth”—turns structural dispossession into a manageable industry story. The isolated court victory encourages attention toward compensation and rights disputes rather than ownership and control of the productive system. It records the wound accurately, then understates the hemorrhage.

The Verdict

China’s short-video and e-commerce sectors are early casualties of the Discontinuity Thesis. If the reported ratios are even approximately correct, human actors, presenters, and much of the production chain are already mechanically obsolete. Social death follows as contracts, status, and income disappear.

The surviving roles are narrow: AI-capital owners, controllers, verifiers, rights gatekeepers, relationship specialists, and infrastructure maintainers. That is not a renewed labor market. It is a shrinking perimeter of servitors around expanding synthetic production.

This text does not alone prove that the entire post-WWII order has already died. It does show one of the mechanisms of its death operating at scale: AI severs productive participation from rising output.