Bengaluru AI Traffic Camera Flags Strapped Guitar as Helmetless Pillion Rider

Bengaluru Traffic Police's automated enforcement system erroneously issued an online traffic challan in mid-September to a scooter rider for pillion riding without a helmet, after an artificial intelligence-powered camera mistook a guitar strapped to his back for a human passenger.

The error came to light when the motorist checked the Bengaluru Traffic Police (BTP) e-challan website and discovered he had been booked for carrying an unhelmeted passenger despite riding alone. A senior traffic police officer confirmed that the mistake originated from a camera under the Intelligent Traffic Management System (ITMS) and was subsequently approved due to human oversight during the secondary verification process.

The incident triggered online debate over the shortcomings and reliability of automated policing tools. BTP officials maintained that combining algorithmic detection with manual checks yields the best operational outcomes, though anomalous bookings can occur.

Bengaluru currently issues contactless penalties across three platforms: ITMS, Field Traffic Violation and Regulation Analytics (FTVR), and the citizen-reporting portal Public Eye. Implemented in December 2022 to free personnel for field regulation, the ITMS uses machine learning to identify infractions automatically, transmitting camera feeds directly to the Traffic Management Centre (TMC) at police headquarters for review before notices are dispatched.

Joint Commissioner of Police (Traffic) Karthik Reddy stated that the ITMS operates at nearly 99% accuracy with minimal human intervention, describing erroneous citations as negligible. According to Actionable Intelligence for Sustainable Traffic Management (ASTraM) records, the BTP issues an average of roughly 26,500 contactless challans daily, with more than 19,000 flagged by AI cameras. Of those automated cases, an average of 25 citations, or 0.13%, are formally disputed each day.

However, another senior officer explained that the system's baseline accuracy drops to around 90% when accounting for errors caught and filtered out by manual reviewers at the TMC. System accuracy had dropped to 80% late in 2023 for specific infractions—such as seatbelt usage, signal jumping, and pillion helmet compliance—before technical updates raised performance over the past two years.

Police sources noted several other recurring false triggers, including motorists wearing black shirts being flagged for driving without seatbelts, font variations on registration plates being classified as defective, and faded zebra crossings prompting unwarranted citations. Commuters pausing at prolonged signals after removing helmets have also been improperly logged by the system.