Waymo Software Flaw Reportedly Leads to Double Collision with Tow Truck

Two Waymo autonomous vehicles reportedly hit the same tow truck under unusual towing conditions due to a software misinterpretation in Phoenix, Arizona. Waymo reportedly issued a software recall and updated its fleet to prevent future such incidents.

Risk classification

• Primary risk domain: 7 AI system safety, failures, & limitations
• Primary risk subdomain: 7.3 Lack of capability or robustness

The Waymo Driver failed to perform robustly when encountering an unusual towing configuration, leading to incorrect motion prediction and subsequent collisions.

Causal factors

• Entity: AI
• Intent: Unintentional
• Timing: Post-deployment

The collisions were caused by the Waymo autonomous vehicle's software incorrectly predicting the motion of a towed vehicle post-deployment, which was an unintentional outcome of its driving operations.

EU AI Act risk tier

• Risk tier: 2 High Risk

High Risk: The AI system is used in critical infrastructure (road traffic and transport), where failures can have significant implications for public safety.

AI system and alleged parties

• AI system: Waymo Driver (Waymo)
• AI purpose: Autonomous Driving; Navigation Assistant
• Behaviour type: Autonomous
• Alleged developer: Waymo, Alphabet
• Alleged deployer: Waymo
• Alleged harmed parties: Unnamed owner of tow truck

Harm severity

Severity ratings are based on this scale.

Highest direct severity in any category: Negligible. Severity is scored from Negligible to Catastrophic in each harm category, for harm the reports describe as caused directly or indirectly by the AI system.

• Physical: direct Negligible, indirect Negligible
• Infrastructure: direct Negligible, indirect Negligible
• Property: direct Negligible, indirect Negligible
• Financial: direct Negligible, indirect Negligible
• Environmental: direct Negligible, indirect Negligible
• Malicious content: direct Negligible, indirect Negligible
• Differential treatment: direct Negligible, indirect Negligible
• Civil rights: direct Negligible, indirect Negligible
• Democracy: direct Negligible, indirect Negligible
• Privacy: direct Negligible, indirect Negligible
• Psychological: direct Negligible, indirect Negligible
• Epistemic: direct Negligible, indirect Negligible
• Child sexual exploitation and abuse: direct Negligible, indirect Negligible

People affected

• Occurrences reported: 2
• People reportedly exposed: 1

Potential causes

Management

• Continuous Fleet Operation: Vehicles continued operating despite the unresolved software failure.

Technology

• Incorrect Motion Prediction: The AI software wrongly predicted how the towed vehicle would move.
• Sensor Input Misinterpretation: Cameras and sensors failed to correctly interpret the angled towed truck.
• Lack of General World Model: The system lacked a conceptual understanding of a backwards-towed vehicle.

Data Inputs

• Unfamiliar Towing Configuration: Training data lacked scenarios with angled, backwards-facing towed trucks.
• Out-of-Distribution Scenario: The persistent orientation mismatch was outside standard training inputs.

Human Factors

• Lack of Human Fallback: No human operator was present to override the AI system decision.

Process and Methods

• Inadequate Edge Case Handling: System struggled to handle novel, complex, and dynamic road scenarios.
• No Real-Time Fleet Shutdown: The fleet was not halted immediately after the first collision occurred.

Information quality

• Classification confidence: High
• Reason for confidence: The reports provide consistent, detailed accounts of the incident from both the developer (Waymo) and independent media outlets, clearly explaining the technical failure, the sequence of events, and the subsequent mitigation steps.

Two Waymo autonomous vehicles in Phoenix, Arizona, collided with the same improperly towed pickup truck due to a software error in predicting the vehicle's motion. The incident resulted in only minor property damage and no injuries. Waymo successfully deployed a software patch to resolve the issue, representing a standard commercial safety event with negligible national security implications.

• Overall national security impact: Negligible
• Scope: Single nation
• Primary target: United States
• Alleged perpetrator: Unknown

Threat characteristics

• Imminence: Long-term. The software issue was successfully resolved via a fleet-wide update, presenting no active or imminent threat.
• Autonomy: Full autonomy. The Waymo vehicles were operating in full autonomous mode, making navigation and motion prediction decisions without human intervention.
• Novelty: Established threat. Navigation errors and collisions involving autonomous vehicles encountering edge-case road scenarios are established risks in autonomous driving technology.

Impact by dimension

• Physical security: Negligible. Minor traffic collisions involving two empty autonomous vehicles and a towed truck resulted in no injuries and only minor bumper damage, posing no threat to national critical infrastructure or physical security.
• Information security: Negligible. No information warfare, intelligence compromise, or disinformation campaigns were associated with this autonomous vehicle navigation failure.
• Sovereignty: Negligible. The incident was a localized traffic event with no impact on state authority, territorial control, or core government decision-making.
• Economic security: Negligible. The software failure and subsequent voluntary recall represent standard commercial product debugging with no threat to strategic industries or national technological competitiveness.
• Societal stability: Negligible. A localized traffic incident involving empty autonomous vehicles does not impact societal stability, civil liberties, or human rights.