Human drivers keep crashing into Waymos

2023-07-27 · Source: Understanding AI · Field: Transportation & Mobility — Autonomous Vehicles & Smart Transportation, Mobility Services & Technology · Depth: Intermediate, long

Summary

An analysis of 78 serious crashes involving driverless Waymo vehicles reported to NHTSA between August 15, 2025, and March 16, 2026, highlights that most incidents, 48 out of 78, were rear-end collisions caused by other vehicles, often while the Waymo was stopped. Another 12 involved other vehicles hitting a stopped Waymo from different directions, and 12 involved collisions with moving Waymos. The analysis details four cases where Waymo's actions might have contributed, primarily due to "excessive caution" leading to abrupt or prolonged stops, such as a vehicle stopping on US 101 due to a routing error, or remaining on a teenager's foot for over eight minutes after an incident. Waymo's overall safety record, with 82% fewer injury crashes compared to human drivers, is noted as strong, with most incidents attributed to human driver error. The article also briefly touches on crashes from other robotaxi companies like Tesla, May Mobility, and Zoox, and the opacity of the Chinese market.

Key takeaway

For autonomous vehicle developers refining system behaviors, this analysis underscores the need to balance caution with dynamic response. Your systems must not only avoid initiating dangerous maneuvers but also intelligently resolve situations like unexpected stops or geofence breaches, preventing secondary human-caused collisions. Prioritize robust fallback strategies that minimize prolonged vehicle immobility in traffic, especially in "no stopping" zones, to mitigate risks from human driver reactions.

Key insights

Waymo's driverless vehicles demonstrate strong safety, with most reported crashes stemming from human driver errors or Waymo's excessive caution.

Principles

• Autonomous vehicle caution can paradoxically increase crash risk.
• Human driver behavior remains a primary AV crash factor.
• AV operational domain limits require robust fallback protocols.

In practice

• Implement dynamic response protocols for unexpected stops.
• Enhance AV perception to anticipate human driver errors.
• Refine geofencing to prevent AVs from entering unqualified zones.

Topics

• Autonomous Vehicles
• Waymo Safety
• Robotaxi Crashes
• Human Driver Error
• NHTSA Reports
• Operational Design Domain

Best for: CTO, VP of Engineering/Data, Director of AI/ML, Robotics Engineer, Policy Maker, Tech Journalist

Related on AIssential

• 99% Car Accidents Are Caused by Human Error — Autonomous systems can achieve superhuman safety performance by eliminating human error in complex tasks like driving.
• Emergency First Responders Say Waymos Are Getting Worse — Emergency first-responder leaders, including city firefighters, police officers, EMTs, and paramedics, conveyed significant frustration to federal regulators last month regarding the performance of au…
• Tesla reveals two Robotaxi crashes involving teleoperators — Teleoperator intervention in Tesla Robotaxis has led to low-speed crashes, revealing safety challenges in autonomous vehicle deployment.
• Behavior Prediction under Occlusion — Accurate, uncertainty-aware behavior prediction is crucial for safe, efficient autonomous navigation.
• Waymo is asking DoorDash drivers to shut the doors of its self-driving cars — Autonomous vehicle operations face unexpected human-interface challenges requiring novel, human-powered solutions.
• On the Road With DRIVE AV — Autonomous vehicles can provide comfortable, confident, and anxiety-free navigation in complex urban environments.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Understanding AI.