Dhaka’s AI Traffic System Cuts Police Quarrels, Issues First $16 Fine

In a city where average traffic speed barely exceeds walking pace, authorities have turned to artificial intelligence to enforce road rules. Dhaka’s streets are choked with buses, cars, motorcycles and pedal rickshaws that often ignore signals, pedestrian crossings and even officers. Before the AI system, on‑the‑spot fines frequently sparked arguments, and police sometimes were struck by vehicles refusing to obey commands. In April, police linked existing traffic‑camera feeds to AI software designed to spot violations such as running red lights, lane infringements and illegal parking. The system flags potential offences for human verification before a fine is issued.

Traffic sergeant SM Nazim Uddin said that since the AI was introduced, drivers have begun obeying traffic laws and police no longer face daily quarrels over violations. Motorist Hannan Rahman Jibon received an automatic 2,000 taka ($16) fine after running a red light; the AI system detected the offence and sent a text alert to the vehicle’s owner. Police spokesman N.M. Nasiruddin said the AI traffic system has led to the prosecution of at least 300 vehicles so far.

The technology shifts enforcement from manual stops to automated detection, reducing direct confrontations between officers and violators. Early data suggest a change in driver behaviour, with many now heeding signals to avoid fines. However, the system still struggles with blurred or small number plates and does not yet cover pedal rickshaws, a major part of Dhaka’s traffic mix. Police plan to expand camera coverage, improve plate recognition and add detection for footpath violations and rickshaw movements in the next phase. Officials acknowledge that technical limits remain, including difficulty reading obscured plates and the need to integrate non‑motorised vehicles. They are working with the road transport authority to improve image quality and to design rules for rickshaws.

Officials will monitor whether the AI‑driven approach can sustain compliance as it scales to more intersections and incorporates additional vehicle types. A pilot expansion to additional districts is slated for the coming months, with performance metrics to be published quarterly.