Robotaxi hits pedestrian, exposing problems during rapid expansion?
Produced by Zhineng Technology
In 2025, we see L2+ assisted driving shifting from equal access to a focus on safety, with the entire industry becoming more cautious.
The Robotaxi accident in Zhuzhou, Hunan at the end of the year, which resulted in injuries, has brought China's potentially rapid advancement of Robotaxi (including RobotVan) back to the core issue: How far is L4-level Robotaxi (including RobotVan) from safe, large-scale commercial deployment?
Two key questions: Why did the accident happen at this point in 2025? Is there a necessary link between "accelerated expansion" and "increased accidents"?
01
Why did the accident happen at this time?
The "rapid start" of new Robotaxi players
What's hidden behind it?
On December 8, the Robotaxi accident in Zhuzhou, which resulted in injuries, was one of the few publicly confirmed Robotaxi accidents in China that caused injuries. Two motorcycle riders were hit, with one dragged under the vehicle, marking a widely noticed Robotaxi injury incident.
About 80 meters north of the intersection of Yanjiang Road and Qiyi Road in Lusong District, around 9 a.m., the weather was clear, the ground was wet and reflective, possibly just after a sprinkler passed. The road had four lanes in both directions, with open visibility and clear lane markings. The accident vehicle eventually stopped on the sidewalk and no-parking zone, with the two injured lying 10-15 meters behind the vehicle. Currently, all accident videos are post-incident footage, with the actual collision moment missing. The truth awaits official and black box data release.
Based on the video and road conditions, we can understand the scene and situation as follows: From the current reconstruction, the Robotaxi was traveling south and collided after approaching the motorcycle.
The motorcycle may have fallen on its own or lost control due to the collision. Subsequently, the vehicle failed to brake in time, and there was a suspected dragging situation.
The current issue with L4-level Robotaxi is the lack of self-correction capability after errors occur. This company entered the market only six months ago and has already deployed about 80 vehicles in a medium-sized city.
Hello is not an ordinary new player, backed by three major capital and resources: Hello, Ant Group, and CATL, all well-known names in China.
It officially announced its entry in June this year and deployed nearly 100 vehicles on 297 kilometers of open roads in less than six months. A few days ago, news announced plans to deploy 50,000 Robotaxis by 2027, marking the beginning of a very aggressive commercial expansion cycle.
Of course, the question arises: Can technical capabilities, data accumulation, and system verification support such a fast pace?
L4-level autonomous driving is a long-term balance of data density, system robustness, and extreme scenario coverage. The faster the implementation, the less safety redundancy; the larger the scale, the more edge cases.
New players started with "large-scale vehicle deployment." When Robotaxi's operational speed outpaces system maturity, how can accidents be avoided?
In reality, the larger the scale, the more likely it is to encounter extreme scenarios.
◎ Baidu's Apollo Go: An airbag deployment accident occurs every 10.14 million kilometers on average
◎ WeRide: Hit by a human-driven vehicle illegally changing lanes in Abu Dhabi
◎ Pony.ai: Encountered multiple traffic accidents, none resulting in casualties
China's Robotaxi products are beginning to expand in scale and operating time, truly integrating into the transportation system. Various "complex scenarios" that we humans cannot anticipate will start to emerge.
Even abroad, Tesla's Robotaxi had its first collision within a month of launch, and Cruise was suspended after an accident where it failed to identify a pedestrian and dragged them several meters.
Only when L4-level systems are truly deployed on a large scale in urban roads will vulnerabilities, misjudgments, and collaboration defects be gradually discovered.
02
02
The issue of injury accidents,
How should responsibility be defined?
As the U.S. promotes the expansion of L4-level operations, while Tesla, Uber, andWaymo are advancing, multiple cities in China are currently promoting Robotaxi licenses, responsibility systems, and insurance frameworks.
There are many issues here, including how to assign responsibility for autonomous driving system accidents, whether backend data can reproduce key decisions, how the insurance system interfaces with AI driving behavior, and how core companies can assume long-term safety responsibilities.
This is not only about facing accidents but also about being responsible for public trust in society (ride-hailing drivers also need to see if AI is up to the task, and passengers need to see if it's truly safe inside).
L4-level autonomous driving is a long-term game of data density, system robustness, and extreme scenario coverage. The faster the implementation, the less safety redundancy; the larger the scale, the more edge cases.
The challenge for Robotaxi is the endless "low-probability, high-uncertainty events" on the road. What we need to determine is whether L4-level technology can ensure redundancy in abnormal scenarios, correct misjudgments in real-time, and self-complete cognition in "extreme long-tail scenarios."
When Robotaxi truly has to "share the road with human drivers," are we ready?
The copyright of this article belongs to the original author/organization.
The views expressed herein are solely those of the author and do not reflect the stance of the platform. The content is intended for investment reference purposes only and shall not be considered as investment advice. Please contact us if you have any questions or suggestions regarding the content services provided by the platform.
