If you didn’t trust self-driving taxis before, you definitely won’t after this

Zoox Robotaxis Recalled Due to Dangerous Maneuvers

Zoox, a company owned by Amazon, has initiated a recall of 332 of its robotaxis. The reason was a software error that caused the autonomous vehicles to unnecessarily cross double lane markings and enter or stop in oncoming traffic lanes. This significantly increased the risk of a collision.

The National Highway Traffic Safety Administration confirmed that in certain situations, especially near intersections, the company’s vehicles unnecessarily crossed lane markings. In some cases, they could come to a complete stop in the path of all traffic.

The Incident That Started the Investigation

The first case that drew attention occurred on August 26. While making a wide right turn, a robotaxi entered an oncoming lane and temporarily stopped in it. This prompted a detailed investigation.

Zoox Robotaxi on the road

The Scale of the Problem and Its Resolution

Between August 26 and December 5, the company recorded 62 similar incidents. It is important to note that none of these situations led to an accident or collision. Zoox immediately began working on a fix.

Causes of the Errors

The recall report states that the analysis revealed several main causes of the incorrect behavior:

Among them are incorrect recognition of double-parked cars, unexpected changes in routing, and unsuccessful attempts to avoid blocking intersections. The company stated that it will continue to monitor the fleet’s operation and make further adjustments to improve the driving algorithms.

Zoox Robotaxi in Las Vegas

This case clearly demonstrates the challenges faced by the autonomous transportation industry. Even in the controlled environment of a city, unforeseen road situations require constant improvement of artificial intelligence. Zoox’s prompt response to the incident—from detection to updating the entire fleet—shows that safety remains a priority, and the processes for recall and error correction are functional. However, for broad public acceptance of such technologies, such incidents must become extremely rare, as user trust is built precisely through stable and safe operation.

Leave a Reply