Dhaka has rolled out its first AI-powered traffic enforcement system to tackle severe congestion and improve compliance with traffic laws, using cameras to auto-detect violations.

The Dhaka Metropolitan Police have deployed over a hundred AI-driven cameras across the city to automatically identify a range of violations and enhance road discipline, with initial coverage at key intersections.

Footage is stored on the central server for review, with notices initially sent by post and plans to add SMS/app alerts and offender video access to boost transparency.

More than 20 cameras are already operational at key intersections, with plans to expand citywide as part of a broader AI surveillance rollout.

Initial enforcement targets the worst offenders while gradually expanding coverage, with manual enforcement winding down over time.

The system analyzes feeds in real time in a control room, with analysts validating offenses before notices are issued to reduce driver-police confrontations.

The system routes camera feeds to a control room where analysts review flagged violations before enforcement, helping reduce confrontations between police and drivers.

Early indicators show hundreds of prosecutions and a noticeable shift toward more cautious driving at major junctions in the initial weeks.

The initiative aims to create a transparent digital record of violations, lowering confrontation with police and nudging residents and travelers toward safer, more predictable road use.

This AI-enabled enforcement fits a wider national push to modernize policing and digital governance, including smart CCTV for real-time threat detection and broader AI applications.

Officials caution that long-term traffic improvements depend on broader reforms like infrastructure expansion and removing unfit vehicles, in addition to technology.

The technology uses existing CCTV feeds and AI to detect violations such as red-light running, lane infractions, illegal parking, and soon footpath driving, with automated logging and eventual fines.