This is a research prototype. The data and analyses are preliminary and not yet validated — we'd welcome your .
A traffic enforcement system in Bath, UK, incorrectly issued a fine to a vehicle owner after an automated license plate recognition camera misidentified text on a pedestrian's shirt as a license plate. The error was discovered when the vehicle owners contacted local officials, who subsequently cancelled the fine.
A Bath resident was wrongly fined by the local officials because an automated license plate recognition camera misread the text on her shirt as a license plate number.
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 AI system failed to perform robustly under real-world conditions, misidentifying text on a pedestrian's t-shirt as a vehicle license plate.
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 incident was caused by an automated license plate recognition AI system making an unexpected classification error after being deployed.
An automated traffic enforcement camera in Bath, UK, misidentified text on a pedestrian's shirt as a license plate, resulting in an erroneous fine. The incident has negligible national security implications, representing a minor and easily corrected administrative error with no hostile intent or systemic impact.
The traffic camera system autonomously scanned the environment, detected text on a pedestrian's shirt, matched it to a license plate database, and generated a fine without human intervention.
High Risk: The system is used in law enforcement and traffic management, which are classified as high-risk applications under the EU AI Act.
The primary function of the traffic camera system is to recognize license plates for automated traffic enforcement.
The report clearly details the sequence of events, the specific error made by the AI system, and the resolution of the incident. There is no conflicting information or significant ambiguity regarding the system's failure.
None
Technology
OCR Misidentification
The camera read t-shirt text 'KNITTER' as a vehicle license plate.
Flawed Object Classification
The system failed to distinguish a pedestrian from a motor vehicle.
Data Inputs
Ambiguous Text Input
The t-shirt text closely resembled the personalized plate 'KN19 TER'.
Human Factors
Lack of Human Oversight
No staff reviewed the computer-generated photo before sending the fine.
Process and Methods
Fully Automated Ticket Issuance
The process generated and mailed citations without any manual validation.
Management
Prioritizing Cost over Accuracy
Automation was deployed to reduce labor costs without ensuring reliability.
AI system involved
Identified by automated extraction; can differ from the alleged developers above.
