A Baidu Apollo Go robotaxi fell into a construction pit in Chongqing, China, while carrying a passenger.

The passenger was uninjured and was helped out by residents using a ladder.

Videos on Chinese social media showed the robotaxi at the bottom of what appeared to be a construction trench.

A local shop owner said barriers and warning signs were present at the site, but it is unclear how the vehicle entered the area.

Reuters confirmed the video’s location but could not verify how the vehicle ended up in the pit.

The incident has sparked discussions on social media about the safety of robotaxis in China.

Baidu operates autonomous taxi services in several cities, including Chongqing, Wuhan, and Beijing, and recently signed partnerships with Uber and Lyft to expand its business.

It follows a May incident where a Pony.ai robotaxi caught fire in Beijing while being handled by service staff.

🔗 Source: Reuters

🧠 Food for thought

1️⃣ Autonomous vehicles consistently struggle with unexpected environmental hazards

The Baidu robotaxi’s fall into a construction pit reflects a broader pattern where autonomous vehicles fail to properly classify unexpected road conditions and obstacles.

The 2018 Uber incident provides a clear example, where the vehicle detected a pedestrian 5.6 seconds before impact but failed to classify her correctly as a pedestrian until it was too late to avoid collision1.

Uber’s system specifically could not recognize jaywalking pedestrians unless they were near crosswalks, demonstrating limitations in handling non-standard scenarios2.

This issue extends across the industry, with 3,979 reported autonomous vehicle incidents in the U.S. as of June 2024, suggesting these classification and detection failures are systemic challenges rather than isolated technical problems3.

The construction pit incident highlights how current autonomous systems may excel in controlled environments but struggle when faced with road conditions that don’t fit their programmed parameters.

Recent Baidu developments