© The Indian Express Pvt Ltd
Tags:
Journalism of Courage
A murder accused from Maharashtra fled to Bengaluru, believing the city’s size would help him disappear. Technology proved him wrong.
The man, allegedly affiliated with a political party, entered the city in October 2025 in one car, switched vehicles twice, changed number plates and kept moving across neighbourhoods to evade arrest. The Bengaluru Police, however, pieced together his movements using AI-enabled cameras and Automatic Number Plate Recognition systems. His movements across multiple junctions were mapped, vehicles correlated, and he was finally arrested within the Konanakunte police limits.
Facial recognition inputs, movement analysis and camera-based tracking led investigators to Basaveshwara Nagar, where he was arrested. City police officials say such cases underline how AI surveillance has transformed Bengaluru into a difficult city for repeat offenders to operate in.
With over 9,000 AI-equipped cameras feeding real-time data into the Command-and-Control Centre, Bengaluru Police now rely on technology not just for post-crime investigation, but for prevention, prediction and inter-state coordination.
AI in tackling crime and law & order
AI cameras have proved particularly effective in street crime detection. In a recent molestation case, where a young woman was assaulted and there were no leads on the accused, officers used camera analytics to determine whether the accused fled on foot or by vehicle, tracking footage across multiple junctions to narrow down escape routes.
Missing persons and missing children’s cases are increasingly routed directly to the command centre. Photographs are fed into the facial recognition system, generating alerts if a match appears anywhere in the city. Several theft and missing cases that once took days to coordinate are now handled in hours, officials said.
Inter-state coordination has also improved, with other state police forces working closely with Bengaluru Police, using shared inputs from AI camera networks to track suspects who cross borders.
New Year’s Eve and Diwali: Preventing chaos before it begins
The preventive strength of AI surveillance was most visible during New Year’s Eve and Diwali — two of the city’s most challenging law-and-order periods.
AI-enabled cameras continuously analysed crowd density and movement in high-footfall areas such as MG Road, Brigade Road, Church Street, Indiranagar and Koramangala. Data from metro stations and streets was converted into colour-coded heat maps — green, orange and red — indicating crowd intensity.
Deputy Commissioner of Police (Crime) Raja Imam Kasim said red zones signalled a risk of stampede-like situations. Based on these alerts, police teams regulated pedestrian movement, diverted crowds and restricted entry into congested areas.
A temporary control room was set up at a hotel on Brigade Road, where live feeds from over 200 cameras installed across key surrounding areas were monitored, enabling police to receive continuous real-time inputs from all major locations. Over 20,000 personnel were deployed citywide, supported by command centre teams analysing AI alerts. Hence, said police, there was no major law-and-order incident, no FIR and no need for force.
During Diwali, light-and-sound detection cameras flagged illegal bursting of firecrackers after a 10 pm deadline. Alerts were instantly relayed to nearby Hoysala patrols, enabling swift, targeted action without large-scale enforcement drives.
Facial recognition cameras
Facial Recognition Systems (FRS) are an important aspect of Bengaluru’s Safe City project. These cameras, installed at sensitive locations, scan faces and compare them with police databases of known offenders. Matches trigger alerts at the command centre, which are cross-verified before action.
Officials stress that FRS functions as a deterrent. During New Year’s Eve celebrations, no high-risk individuals were detected, reinforcing both public safety and confidence.
AI in traffic enforcement
AI surveillance is also reshaping traffic policing. Automatic Number Plate Recognition cameras automatically detect violations, track stolen vehicles and assist accident investigations. In the week around New Year’s Eve, police filed nearly 5,000 drunk driving cases, aided by AI-based alerts and targeted deployment.
Senior officers say the visible presence of AI-backed enforcement has acted as a deterrent, contributing to a period with no major accidents or crowd-related FIRs.
“These systems automatically detect violations across key junctions, allowing us to issue challans contactlessly and reallocate our personnel to manage traffic flow and road safety more proactively,” said a police officer, highlighting that artificial intelligence and machine learning help record and process violations with minimal human intervention.
Use of ‘Magic Box’
A key innovation supporting all these efforts is the use of ‘Magic Boxes’. Under the ‘Magic Boxes’ system, the city police temporarily install CCTV cameras at different locations, along with integrating CCTV cameras from apartments and commercial establishments via IP addresses, during major events, vastly expanding coverage without permanent installations.
By analysing crowd behaviour, movement patterns, sound, facial data and traffic flow in real time, Bengaluru’s AI camera network has turned surveillance into an active crime-fighting ecosystem, preventing incidents before they begin, and ensuring the city’s public spaces remain safe.
