This is a research prototype. The data and analyses are preliminary and not yet validated — we'd welcome your .
Amazon's Zoox and Alphabet's Waymo recalled autonomous vehicles following incidents where the self-driving cars interfered with emergency responders, blocked emergency vehicles, and failed to recognize safety conditions like smoke and traffic cones.
On June 20, 2026, an unoccupied Zoox autonomous vehicle reportedly entered an active fire scene after heavy smoke obscured its surroundings. The purported automated driving system braked hard while attempting to steer away and stopped before the vehicle was reversed under teleguidance. Zoox later recalled 105 vehicles and issued a software update intended to improve detection of and response to heavy smoke.
Domain classification, causal taxonomy, severity scores, and national security assessments were LLM-classified and may contain errors.
AI systems that fail to perform reliably or effectively under varying conditions, exposing them to errors and failures that can have significant consequences, especially in critical applications or areas that require moral reasoning.
The autonomous vehicles failed to perform reliably under challenging environmental conditions, such as heavy smoke, and failed to recognize standard safety indicators like flashing lights and traffic cones.
AI system
Due to a decision or action made by an AI system
Unintentional
Due to an unexpected outcome from pursuing a goal
Post-deployment
Occurring after the AI model has been trained and deployed
The safety risks were caused by the autonomous driving systems failing to correctly detect and respond to emergency scenes, which was an unexpected and unintended outcome of their post-deployment operation.
Autonomous vehicles from Zoox and Waymo caused localized disruptions to emergency responders in the United States, leading to federal investigations and software recalls. The incidents highlight critical safety and robustness challenges in integrating autonomous systems with public infrastructure, posing minor physical security and operational risks but no immediate threat to national security.
The self-driving vehicles operate independently in real-time traffic environments, responding to sensory inputs and making driving decisions without human drivers, though they can receive remote teleguidance when stuck.
High Risk: The AI systems are used in critical transport infrastructure and safety components of vehicles. The reports describe robotaxis failing to safely interact with first responders, which poses significant safety concerns to the general public.
The primary goal of the AI systems in both Zoox and Waymo vehicles is autonomous driving, navigating public roads safely without human intervention.
The reports from Reuters and Autoweek provide consistent, factual accounts of the Zoox recall and specific incidents involving both Zoox and Waymo vehicles. The details regarding the June 20 incident, the regulatory response from NHTSA, and the scope of the recalls are clearly documented with direct quotes from officials.
None
Technology
Inadequate Smoke Detection
Software failed to detect and respond properly to heavy smoke obscuring path.
Failure to Detect Warning Signals
AV software failed to recognize flashing lights, flares, and traffic cones.
Data Inputs
Obscured Sensor Data
Heavy smoke degraded sensor inputs, preventing accurate scene assessment.
Human Factors
Reliance on Remote Operator
Vehicle required manual intervention via teleguidance to reverse and exit.
Process and Methods
Inadequate Interaction Protocols
Lack of robust processes for AVs to safely interact with first responders.
Management
Incomplete Risk Assessment
Management failed to fully address emergency scene edge cases before deployment.
Regulatory Environment
Lack of Standardized AV Rules
Absence of clear federal standards for AV interaction with emergency scenes.
AI system involved
Identified by automated extraction; can differ from the alleged developers above.
