or years, millions of delivery riders nationwide have been caught between traffic laws and the delivery clock. Time spent at red lights counted against their delivery window, an invisible tax on their income. Now the clock stops while riders wait.
Suzhou in Jiangsu Province was the first to implement it on July 31. On August 13, Beijing announced a citywide rollout after holding a dialogue with major platforms and their riders. By August 17, Wuxi had become the first city to achieve full coverage, connecting all 4,253 intersections with traffic signals.
The effort on strengthening rider protections through more reasonable delivery algorithms is scaling from pilot to nationwide practice. In Beijing, e-bike delivery times will now assume an average speed of no more than 15 km/h, more time will be added for bad weather and rough roads, orders delayed by slow restaurant prep will no longer hurt a rider's service score and merchant-page delivery estimates will be set more realistically. Together, the changes pull delivery windows back toward a realistic estimate to account for human limits and conditions riders cannot control.
The measures address a specific problem, and it offers a useful lens on something larger and far harder to legislate away: the concept of neijuan, or involution. It is a term that has, in just a few years, migrated from academic sociology into complaints about excessive competition in the Chinese economy and in everyday life.
Involution describes a particular kind of "race-to-the-bottom" competition, in which everyone works harder, yet overall welfare fails to grow, and can even shrink. Students study later into the night to hold the same class rank. Employees compete to stay at their desks past the end of the workday, until everyone ends up working longer hours. Factories slash prices below cost to defend a market share that keeps shrinking anyway. And, in this case, delivery riders run red lights to shave seconds off a delivery window set by an unrealistic algorithm at the cost of everyone's safety on the road.
In recent years, authorities have repeatedly vowed to address involution, and their tactics have evolved. In 2021, facing complaints that after-school tutoring was forcing children into ever-longer study hours just to keep pace with classmates, the central government reached for the bluntest available tool and banned for-profit tutoring in core school subjects outright, a move that reshaped an entire industry overnight, though private one-on-one lessons and unlicensed online tutors have resurfaced recently.
The measures now targeting delivery-rider algorithms point to a different approach: instead of a sweeping campaign, authorities identified the specific mechanism distorting behavior, then negotiated a solution with the platforms and workers who depend on it.
The new measures, though limited in scope, shed an encouraging light on how China might address involution. For years, it has been treated as a work-culture problem deeply embedded in society, and the authorities' response has largely been limited to campaigns whose effects rarely last. This model suggests that seemingly intractable cultural problems often stem from technological design that can be fixed once the people affected are brought into the decision-making process.
It also shows that in an era of advancing technology and AI, the tools that caused the harm can be turned toward fixing it. Rather than resorting to slogans and blanket mandates, the government should replicate this model across sectors, chipping away at involution mechanism by mechanism, industry by industry, through targeted, negotiated fixes.
Old Version
