Author | Lin Siyuan, Editor | Zhang Lianyi
Recently, an incident in Shenzhen where a passenger's hand was clamped by an autonomous vehicle has sparked public attention. Claims such as "the door was locked and could not be opened" and "the passenger's finger was severed" spread rapidly, pushing public sentiment toward questioning the safety of autonomous driving.
However, as the facts are restored one by one, we find that this is an accidental incident caused by the passenger's manual operation, which has nothing to do with the autonomous driving technology itself. On this basis, a more worthwhile question to discuss is—how should we approach a technology that is still growing?
01. Fact Clarification: What Exactly Happened?
Through frame-by-frame restoration of the video, several key facts need to be seen:
The door was not locked. The video shows that at the time of the incident, the door handle of the left rear door of the involved vehicle had already popped out and was in a normal openable state. The online rumor that "the door was locked" does not match the facts. In fact, the passenger could have escaped by simply raising their hand and pulling the left rear door. According to the on-site video, the passenger may have been nervous after their hand was clamped, mistakenly assuming it was the driver's door that clamped their hand, and did not attempt to pull the left rear door handle, resulting in a failure to escape in a timely manner.
"The driver's door cannot be opened from the outside" is a design feature, not a malfunction. Since the driver's seat of an autonomous vehicle is unoccupied, restricting external opening is to prevent passengers from mistakenly entering the driver's seat, which is a common safety design in the industry.
At the time of the incident, the vehicle was stationary and picking up passengers, and the autonomous driving system was not performing any driving tasks. The direct cause of the hand clamping was that when a fellow passenger was closing the door, the hand of the person assisting the boarding happened to be in the door closing area—this is a purely human operational negligence. After detecting the anomaly, the vehicle did not move either, causing no secondary injury.
The injured person was examined and found to have no serious injuries, and has returned home to rest. The online descriptions of severe injuries are not true.
In short: this is not a technical accident, but an operational mishap. In the taxi and ride-hailing industry, accidents of passengers getting their hands clamped by doors are not uncommon.
After the incident, the involved company, Pony.ai, issued a statement at the first time, clarifying that the incident was caused by the passenger's improper operation and had nothing to do with the vehicle's safety design and autonomous driving capabilities.
After reviewing the complete video, the Shenzhen Traffic Police also determined that the incident was an accidental event and did not constitute a traffic accident.
02. How Should We Treat a Technology That Is Delivering on Its Promises?
The original intention behind the birth of autonomous driving technology is precisely to solve the most fatal shortcomings in human driving—negligence, distraction, fatigue, and misjudgment. More than 1.2 million people die in road traffic accidents globally every year, over 90% of which are caused by human factors. The ultimate value of autonomous driving is to replace uncertain human operations with deterministic system capabilities.
This value is being validated by data:
As a frontier city in the application and implementation of autonomous driving technology in China, Shenzhen has multiple companies operating Robotaxis, and the overall liability accident rate per 10,000 kilometers is far lower than that of traditional taxis and ride-hailing services.
As of now, the total global operating mileage of leading domestic autonomous driving companies has exceeded the 100 million kilometers level, and the statistical safety performance is superior to that of human drivers.
Of course, this does not mean that autonomous driving will not encounter problems. Technology implementation is a process of continuous iteration in the real world, and safety details in the operational links also need to be continuously optimized. In fact, safety prompts for the boarding and alighting process—such as in-vehicle voice broadcasts and passenger-side page reminders—are being improved and promoted within the industry.
But there is a bottom line that must be held: it is unfair and harmful to deny the systematic safety value of a technology using sporadic human operational accidents.
Accidents such as hands getting clamped when closing doors and pedestrians being hit when opening doors happen every day with traditional taxis and ride-hailing services, but we would not deny the entire taxi industry because of this. The same understanding and rationality should also be given to autonomous driving. And for the autonomous driving industry to win this understanding, it also needs to take a more proactive stance and continuously refine every safety detail.
03. Conclusion
Every technology that changes human mobility has moved forward amidst skepticism. This was the case when automobiles replaced horse-drawn carriages, when civil aviation was commercialized, and it is the same with autonomous driving.
Safety is the bottom line of this industry, and also its starting point. The stringent demand for safety should be directed towards stricter standards and more solid technology, rather than towards fear and misunderstanding.
Faced with an accidental event that has been determined by facts and official authorities, what we need is not an emotional trial, but fact-based judgment, data-based trust, and the due patience for technological civilization.
Do not let a single accident obscure the safety bottom line that an industry is guarding.