Bengaluru: AI cameras issuing wrongful fines despite wearing seatbelts? Read this

AI cameras in Bengaluru, designed to monitor traffic violations, have faced issues with incorrect seatbelt violation fines due to colour confusion. A recent incident highlighted errors where the seatbelt colour matched the driver's shirt, leading to wrongful fines. Similar problems with AI cameras have been reported in other states.

Vinaykumar Patil

First Published Aug 19, 2024, 2:09 PM IST | Last Updated Aug 19, 2024, 2:09 PM IST

In Bengaluru, AI cameras installed at traffic signals and along major roads, including the Bangalore-Mysore Expressway, are designed to monitor traffic violations such as speeding and seatbelt use. These cameras are meant to function like vigilant eyes of the police, capturing infractions and automatically sending penalty challans to offenders. However, recent incidents reveal that these AI cameras are making significant errors, particularly with seatbelt violations.

A recent post on Reddit shared a troubling experience of a driver fined by an AI camera despite wearing a seatbelt. The driver’s seatbelt was the same colour as his shirt, leading the camera to wrongly identify him as not wearing one. This error occurred twice, as shown in images from the camera, which were incorrectly used to issue two fines for the same alleged violation. The seatbelt’s colour matched the driver’s shirt, confusing the AI’s detection system.

AI-equipped vehicles to cover 1400 km weekly for monitoring 'Brand Bengaluru' issues

In India, most cars come with black seatbelts, but premium models often feature beige or lighter-coloured belts to match the car’s interior. Unfortunately, the AI cameras seem ill-equipped to detect these variations, resulting in wrongful fines. This is not an isolated issue; similar problems have been reported in other states.

AI cameras capture over 12,000 traffic violations on Bengaluru-Mysuru expressway within 16 days

In Kerala, over 700 AI cameras were deployed to curb road accidents. However, they faced problems such as misreading registration numbers and confusing rivets or screws on number plates with zeroes. One notorious incident involved an AI camera incorrectly recording a bike’s speed at 1240 mph, leading to a fine for not wearing a helmet, even though the biker was not at fault. This fine was eventually cancelled after further review.

The intent behind installing AI cameras is to automate traffic enforcement, reducing human error and increasing efficiency. These cameras are supposed to capture violations, which are then reviewed by officials before issuing fines. Despite their high-tech capabilities, the initial rollout faced several issues, leading to wrongful fines and administrative challenges. Authorities have since introduced additional oversight, with enforcement officers now verifying images before fines are issued.