The "Zero Fault" Myth Shattered: Radish Express Stops Running, Exposing the Deadly Weakness of Autonomous Driving

When it comes to stock trading, rely on the Golden Qilin analyst research reports—authoritative, professional, timely, and comprehensive—to help you uncover high-potential themes and opportunities!

On the evening of March 31, just before 9 p.m., in Wuhan, nearly a hundred “Luobo Kuaipao” driverless vehicles nearly simultaneously powered down. They came to a halt side-by-side in the middle of the elevated express lanes. Some vehicles lined up abreast and blocked two lanes solid; the navigation map turned entirely red, and traffic was close to a complete shutdown. Even more dangerous was that although passengers could open the doors manually, they were trapped in the middle of the fast-moving highway traffic—unable to walk away and unwilling to get out hastily.

The in-car SOS emergency calling system malfunctioned; the customer service hotline was either busy or, once connected, only mechanically repeated “network anomaly.” After being trapped for nearly two hours, people ultimately relied entirely on traffic police officers who walked onto the elevated roadway to guide and evacuate them vehicle by vehicle. In the early hours of the next day, Wuhan traffic police reported it as a “system failure.” Fortunately, there were no injuries or fatalities.

But this incident—almost a citywide traffic disruption—shatters the “zero accidents, zero human intervention” myth that Luobo Kuaipao has long promoted. Under the glow of technological hype, it reveals a harsh reality: so-called Level 4 autonomous driving has not truly rolled out even the most basic fail-safe safety mechanisms.

The problem is not an occasional bug, but a major flaw in the system’s design logic.

Real high-level autonomous driving must have “fail-safe” capability: if the main system collapses, the local backup module should immediately take over and execute a minimum-risk strategy—slow down, flash double hazards, and slowly pull over and park. This is industry consensus and also a safety bottom line. But Luobo Kuaipao’s vehicles directly “brain-dead,” stopping in the middle of the driving lanes, indicating that its architecture heavily depends on cloud instructions or centralized dispatching and lacks independent local emergency logic. Once communications are interrupted, servers malfunction, or an OTA update introduces a compatibility issue, an entire batch of vehicles could go offline at the same time.

Even more worth worrying about is “the synchronized shutdown of a hundred vehicles.” If it were a single-point hardware failure, it wouldn’t spread so widely. If it originated from software pushes, network partitioning, or erroneous instructions from the control center, it exposes the fragility of a centralized architecture—if one node fails, the entire network goes down. This is no longer a minor hiccup in technological iteration; it is a systemic risk that must be addressed before large-scale deployment.

Operationally, it also drops the ball. Data shows that in Q4 2025, the order volume of Luobo Kuaipao reached 3.4 million orders—far beyond the scope of “testing,” entering a quasi-commercial stage. In theory, a scale like this must be paired with dedicated road rescue teams, a 7×24-hour emergency response center, and real-time linkage mechanisms with traffic management departments. But the reality is: after a fault occurs, the company has no on-site personnel arriving first, and no effective remote intervention tools—so it can only wait for public police forces to provide backup.

At its core, this is a form of cost shifting—companies enjoy the low labor-cost benefits brought by “unmanned operations,” while dumping all safety risks onto society.

The deeper issue behind it is a serious disconnect between business timing and institutional development.

As one of the cities with the longest open autonomous driving testing roads nationwide, Wuhan provides Luobo Kuaipao with an almost ideal policy testing ground. The platform leverages this advantage to rapidly win users’ minds with subsidy strategies far below market price, but it did not simultaneously build a responsibility framework that matches it. Who is held liable for accidents? How are claims handled for insurance? How are violations determined? Current regulations still have a blank spot regarding “AI drivers.” This time, fortunately, no one was injured. However, on the elevated roadway of Dongfeng Avenue, a stationary Luobo vehicle was rear-ended by a Tank 300 from behind, causing severe damage to the latter’s chassis—if it had resulted in personal injury or death, the chain of responsibility would fall into a quagmire.

Jiege believes the public is not opposed to technological progress, but refuses to be free test subjects. “I don’t dare to ride again” is not conservatism; it’s rational self-protection. When a company treats public roads as a low-cost validation ground yet cannot even provide safety solutions for the worst case, the collapse of trust is only a matter of time.

This shutdown is a wake-up call: safety is not an added feature of autonomous driving—it is an indispensable prerequisite. Technology can move fast, but deployment must be steady. True intelligent mobility is not about how advanced the algorithms are or how low the prices are; it’s about whether, when things go out of control, it can keep people alive and get them off the vehicle safely. Otherwise, all slogans that “the future is already here” are nothing more than illusions built on shifting sand.

A massive amount of information and precise analysis—right on Sina Finance APP

责任编辑：张乔松