Software Update Initiated for Entire Fleet

Zoox, the autonomous vehicle subsidiary of Amazon, has initiated a voluntary recall of its entire fleet of 105 robotaxis. This move comes as a direct response to a technical incident where one of the company's self-driving cars struggled to navigate through dense smoke surrounding an active emergency scene.

According to an official statement provided to the media, the company is deploying a software patch designed to refine the vehicles' perception systems. The update specifically

“enhances the existing capability of detecting active emergency scenes by adding the ability to detect and respond to heavy smoke in certain situations.”

The Incident in Las Vegas

Details filed with the National Highway Traffic Safety Administration (NHTSA) clarify that the event occurred on June 20 in Las Vegas, where Zoox currently operates a public shuttle service. The robotaxi, which was operating without passengers, encountered a plume of heavy smoke that obscured a fire emergency site not yet marked by traffic cones.

The vehicle’s reaction was erratic: it performed a hard braking maneuver and attempted to steer away before coming to a complete halt. A remote human operator eventually intervened, taking control of the vehicle to manually reverse it away from the area, which allowed first responders to secure the site.

Broader Concerns Over Emergency Response

This event highlights a recurring challenge for the autonomous vehicle industry: maintaining safe operation in unpredictable, high-stakes environments. Critics and regulators have grown increasingly concerned regarding how robotic fleets interact with public safety officials.

The NHTSA has recently expressed frustration over similar incidents involving various companies, including:

  • Waymo vehicles obstructing roads in San Francisco during a power outage.
  • A Waymo robotaxi delaying an ambulance responding to a mass shooting in Austin, Texas.

The NHTSA administrator Jonathan Morrison addressed the industry directly, labeling the current inability of these systems to interact safely with emergency personnel as an “unacceptable” functional failure. “To state it bluntly: an autonomous vehicle that cannot safely interact with first responders is a danger to the general public,” Morrison noted in a formal communication to the sector, emphasizing that emergency scenes should not be treated as rare exceptions in the testing of these technologies.