This is a research prototype. The data and analyses are preliminary and not yet validated — we'd welcome your .
On December 29, 2019, a Tesla Model 3 operating on Autopilot collided with a parked fire truck on Interstate 70 in Indiana. The fire truck had its emergency lights activated while responding to a separate incident. The collision resulted in the death of a passenger and serious injuries to the driver. Subsequent investigations by the NHTSA confirmed the vehicle was in Autopilot mode at the time of the crash, leading to broader regulatory scrutiny of Tesla's driver-assistance systems and their ability to detect stationary emergency vehicles.
A Tesla on Autopilot mode failed to see a parked fire truck and crashed into its rear on an interstate in Indiana, causing the death of an Arizona woman.
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.
Also classified as
Drivers frequently overrely on the Autopilot system, failing to maintain active oversight and intervene when the system fails to detect hazards.
The Autopilot system failed to perform reliably under normal driving conditions by failing to detect a stationary emergency vehicle with flashing lights.
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 collision occurred post-deployment due to the Autopilot system's unintentional failure to detect and react to a stationary emergency vehicle.
Added categories peak at 2: Minor (Psychological Harm, Inferred)
A consumer Tesla Model 3 operating on Autopilot collided with a parked fire truck in Indiana, resulting in one fatality and one injury. While highlighting critical safety limitations of consumer driver-assistance AI and recurring issues with emergency vehicle detection, the incident is a regulatory and public safety issue with negligible national security implications.
The Tesla Autopilot system operates as an Autonomous system, maintaining speed and lane centering by responding to real-time environmental inputs from cameras and radar without high-level planning. The reports note it can brake, accelerate, pass other vehicles and stay within lanes automatically.
High Risk: The system is used in critical transport infrastructure and vehicle safety. The reports describe an Advanced Driver Assistance System (ADAS) that controls dynamic driving tasks, which directly impacts physical safety and human life.
The primary function of Tesla Autopilot is to assist or perform autonomous driving tasks such as lane keeping, speed maintenance, and traffic-aware cruise control.
The reports provide consistent, detailed accounts of the Cloverdale crash and the broader NHTSA investigation. Key facts, such as Autopilot engagement, the nature of the collision, and the resulting casualties, are verified by police and federal safety regulators.
One could argue the primary cause was driver inattention (human error) rather than AI failure, but the NHTSA investigation specifically focuses on the Autopilot system's failure to detect stationary emergency vehicles.
Technology
Autopilot OEDR Failure
Autopilot failed to detect and react to the stationary fire truck in its lane.
Inadequate Static Object Detection
Radar and camera systems struggled to identify stopped emergency vehicles.
Lack of Emergency Light Detection
Autopilot lacked initial ability to recognize flashing emergency vehicle lights.
Data Inputs
Sensor Input Limitations
Radar and camera data inputs were insufficient to resolve stationary obstacles.
Missed Visual Cues
Cameras failed to process scene control measures like cones or flashing lights.
Human Factors
Driver Inattention
Driver failed to perform primary object and event detection during autopilot.
Overreliance on Autopilot
Driver assumed the system would handle braking and hazard avoidance.
Process and Methods
Inadequate Driver Monitoring
Methods to enforce driver engagement during Autopilot use were insufficient.
Insufficient ODD Limitations
Autopilot allowed operation in conditions where its OEDR capabilities failed.
Management
Inadequate Risk Assessment
Tesla deployed Autopilot without fully mitigating stationary hazard risks.
Post-Incident Software Updates
Safety updates for emergency lights were deployed only after multiple crashes.
Regulatory Environment
Lack of ADAS Safety Standards
No strict regulatory standards existed for Level 2 driver assist systems.
AI system involved
Identified by automated extraction; can differ from the alleged developers above.
