Delhi Enables AI-Powered Camera Network Under Safe City Project: Here’s What’s Coming

Starting October 1, more than 3,500 AI-integrated surveillance systems will be activated in Delhi, marking the first step in the city’s ambitious Safe City Initiative. This comprehensive rollout includes the implementation of AI cameras, gunshot detection systems, public warning systems, and urban intelligence systems for proactive information-based policing. It focuses on integrating technology within public safety systems—aimed at averting anticipated threats, streamlining the handling of emergencies, and revolutionizing policing in the capital of the country.

Let’s analyse the practical implications of this surveillance transformation and how advanced technologies are altering law enforcement within the city.

A Tech-First Policing Model: What’s Being Deployed

The first phase focuses on situational awareness and real-time responsiveness with the addition of AI-powered systems. Unlike conventional CCTV cameras, these units form part of a smart surveillance network that actively aids in the collective decision-making process.

ALSO READ: JioPC Launched: Can This Be The Future Of Computers?

Key Components in Phase 1:

AI Video Analytics: From Passive Footage to Actionable Intelligence

The AI-powered surveillance video analytics (AIVA) systems that support Sundance technologies are equipped with capabilities of can identify numerous high-risk behaviours. Some of these are:

• Gunfire or sudden loud noises
• Vehicle theft attempts
• Detection of abandoned objects
• Crowd formation in non-designated zones
• Crimes against women (e.g. mobbing or distress gestures)

As soon as an abnormal event is detected, the C4i command centre at Delhi Police HQ receives real-time alerts. Deputy Commissioners, as well as nearby police stations, also receive alerts simultaneously. This ensures that responses are acted upon from multiple levels in the response hierarchy.

Prioritising Women’s Safety with Real-Time AI Response

A woman's proactive protection is one of the key design principles of the Safe City system, an integrated part of the system that aims to provide. The protective AI recognises high alert distress filtration of women, like visibly or being surrounded/encircled mark, even when no complaint is registered.

High-Priority Zones:

• Women’s hostels and PGs
• University campuses
• Workplaces with large female staff
• Previously mapped hotspots for harassment, snatching, and eve-teasing

Real-time high-priority alerts with automated patrol assignments are issued for high-importance incidents like patrolled zone breaches.

Gunshot Detection and Vehicle Tracking

The system also integrates gunshot detection technology. For triangulation, three acoustic sensors are needed. PTZ cameras also have the capability to auto-zoom to areas of interest. Their footage can be streamed live to police units for situational assessment.

ANPR systems in particular mark and track flagged blacklist vehicles on given checkpoints in real time. Face recognition algorithms scan live footage to identify persons flagged and stored in a public database of more than twenty thousand criminals.

Smart Crowd Monitoring

AI cameras are also trained to detect public and private gatherings. Once the threshold of fifty participants is surpassed, alerts are issued. Cluster formations are instead recorded as low-priority events, but face continuous surveillance for potential escalation.

ALSO READ: OpenAI Delays Launch of Open-Weight AI Model Citing Safety Concerns

This function seeks to prevent flash protests and large-scale disturbances through active predictive policing based on pattern recognition.

Real-Time Verification and Dispatch

Here’s how the response flow works:

1. AI triggers an alert based on incident detection

2. C4i operators verify footage and label it with a priority tag

3. Local teams are dispatched based on location and severity

4. False positives are logged with reasons, building a feedback loop for algorithm refinement

All alerts are accompanied by cloud-stored legal evidence and internal review footage for review, ensuring high-definition, time-stamped video is always available.

Force Multiplier for Police, Not Just a Surveillance Grid

Real-time accountability and evaluation is now possible for officers, enabling supervisors to monitor and assess tactical and crowd management decisions. This provides avenues towards policy change and thorough training evaluation in addition to transparent auditing of public safety events.

“This isn’t just about adopting technology—it’s about transforming how we police the city,” said a senior officer involved with the rollout.

What’s Next in the Safe City Rollout?

During a post-incident audit following the Kanjhawala case in 2022, over 10,000 cameras were identified. Currently, only a small portion of this activation will be implemented, serving as a precursor to a comprehensive review of urban surveillance gaps in Delhi.

The rollout is scheduled to complete in phases until end-2026, with high-risk and densely populated zones given top priority.

Final Thoughts

The installation of cameras in the Capital did not emerge from impulse only. The implementation of the capital’s Safe City project signifies a move toward proactive, smart policing. It integrates real-time analytics, threat detection, facial recognition, and coordination centre interfacing to empower and digitally reinforce the police force to prevent crimes proactively. The start of phase 1 marks the comprehensive modernisation of public safety in the capital, which is set for October this year.