AINEWS
China's AI rollout is hitting ride hailing drivers and creative workers first, and Beijing's twin role as deployer and absorber is running into limits.
"I spent six gruelling years working my way up to become a cinematographer, only to be replaced by a piece of software."
Tian Zemin, a Beijing-based cinematographer with roughly two decades on Chinese film sets, posted that line in July. Recent court rulings in China have begun to award compensation to white-collar workers laid off and replaced by AI, according to Guardian reporting on the case. Tian's post is the kind of human signal that is starting to crowd the edges of a much larger number: an estimated 320 million Chinese workers, about 44% of the country's workforce, now sit in flexible employment, the labor category AI is hitting first. The figure is up from 160 million in 2019, according to one thinktank cited in the same reporting. The institution is unnamed, so the number should be read as directional, not audited.
In March 2026, Baidu's Apollo Go robotaxi fleet in Wuhan suffered a mass-paralysis incident that left driverless vehicles stranded with riders for hours and took the fleet off the streets for months of investigation. Wuhan taxi driver Yao Xinnong, 45, told the Guardian his wages had fallen roughly 40% since Apollo Go launched in the city in 2022. "I don't really like them, to be honest," he said. "Their impact has been huge." On July 23, China quietly resumed issuing robotaxi licenses after a months-long freeze, a signal that deployment pressure on the labor market is back on.
Two labor markets, two policy gaps. Ride-hailing drivers like Yao are facing wage compression from a state-licensed driverless fleet whose operator, Baidu, is one of the country's most systemically important AI platforms. Creative workers like Tian are being individually replaced by generative tools, one contract at a time, with no platform to point at. The first is a deployment story, the second is a substitution story. They share a country and a press cycle, and that is about it.
Beijing has noticed. At the World AI Conference in Shanghai on July 17, President Xi Jinping told the room that AI should be "an important driver for shared prosperity and common security." The framing is policy, not measurement. It commits the government to a political claim about who absorbs the cost of the rollout without naming a single instrument (hiring quotas, retraining scale, replacement taxes) that would actually move the number.
The most likely instrument is the one China has used before. Where the United States lets the labor market take the hit and absorb the politics, Beijing could instruct state-linked employers and platform companies to comply with hiring quotas, the same way it instructed property developers to slow down in 2021. The tool exists; the political appetite is the question. "Shared prosperity" reads better in a World AI Conference keynote than it does in a Wuhan taxi company's payroll.
The cold math underneath all of it: China's state capacity is the reason its AI rollout is the fastest in the world, and it is also the reason the political cost of doing nothing is concentrated in one place. The next eighteen months will show which way that double role tips. If the 320 million figure climbs toward 350 million without a quota, retraining program, or replacement tax, the line Xi delivered in Shanghai will start to read less like a policy and more like a description of a problem the state is letting run.