AI-run cameras covering motor vehicles, but who tames Dhaka’s wild autos?

Highlights:

• AI cameras detect Dhaka traffic violations, filing hundreds of automated cases
• Unregistered autorickshaws evade enforcement due to missing database records
• Illegal vehicles frequently ignore signals, causing congestion at major intersections
• Motorcyclists increasingly follow rules fearing AI-monitored traffic fines
• Experts say regulation must accompany AI for effective traffic management
• DMP preparing policy framework targeting unlicensed vehicles and illegal rickshaws

At a traffic signal in Banglamotor in the capital, the red light is on. A queue of motorcycles and private cars waits behind the stop line.

At the same time, several battery-run autorickshaws coming from the opposite direction edge forward, cutting across the zebra crossing.

Above them, an AI-powered camera keeps rotating, quietly monitoring the intersection.

Across major traffic junctions in Dhaka, the Dhaka Metropolitan Police (DMP) has rolled out automated case filing using artificial intelligence (AI)-based cameras to detect traffic violations. Several hundred cases have already been recorded for offences such as running red lights, crossing stop lines, helmetless riding, and wrong-way driving.

But on the ground, the system tells only part of the story. While it works effectively for registered motor vehicles, a large unregulated segment –  battery-run autorickshaws, number plate-less small vehicles and illegal rickshaws – largely remains outside its reach.

The gap that cameras cannot close

Traffic sergeant Rakib, stationed at Banglamotor, told The Business Standard that enforcement against such vehicles remains difficult due to a lack of registration. He said even when violations are captured by AI cameras, cases cannot be filed because these vehicles are not in the database.

He added that unregulated vehicles are making traffic management increasingly difficult and called for a structured regulatory framework.

During about one hour of observation at Shahbagh, Hotel Intercontinental intersection, Banglamotor, Karwan Bazar and Farmgate, battery-run autorickshaws were seen crossing stop lines, occupying zebra crossings and entering wrong lanes, contributing to congestion.

In contrast, many motorcyclists were seen following signals and wearing helmets. Several riders said fear of fines has increased since the introduction of AI-based monitoring.

At Farmgate, ride-sharing driver Md Rubel told TBS that compliance among motorcyclists has improved due to cameras, but autorickshaws continue to violate rules. He said rule-following riders are often blocked by vehicles moving against traffic.

Traffic officials say the AI system relies on number plates and vehicle databases, making enforcement against unregistered vehicles difficult.

A DMP traffic official, requesting anonymity, said notices can be issued only where ownership data exists. Battery-run autorickshaws without registration cannot be processed for automated case filing even when detected by cameras.

Autorickshaw drivers, however, argue that road conditions force their behaviour. Shahid Mia, a driver in Banglamotor, said it is often difficult to move straight due to pressure from larger vehicles and that passengers frequently request unscheduled stops.

Technology without structure

Prof Md Hadiuzzaman of Buet said AI-run cameras can only do their job effectively if basic vehicle discipline is in place first. He said unregulated autorickshaws and small vehicles must be brought under a defined system, otherwise AI-based traffic management will struggle to deliver results.

He added that AI systems depend on structured data from registered vehicles, and the presence of unregulated traffic "throws a wrench in the system" by weakening enforcement. He stressed the need for a classification-based regulatory framework, noting that cameras alone cannot restore order without stronger structural control.

Experts say AI cameras alone are not enough to bring discipline on the roads and must be backed by tighter enforcement against illegal vehicles, improved registration systems, lane discipline and consistent field-level monitoring.

DMP Additional Commissioner (Traffic) Md Anisur Rahman told TBS that around 500-600 cases were recorded in the first week of AI camera operations. He said a plan is underway to bring unlicensed vehicles, especially rickshaws, under control, and a policy is being prepared.

He added that legal enforcement will improve once the policy framework is finalised, given current challenges in identifying such vehicles.