Mint explainer | Can technology bridge India's public-security gaps after the Delhi Red Fort blast?

How did explosives slip through the cracks?

The improvised explosive device (IED) used in the Delhi Red Fort blast has been reported across platforms to be a mix of ammonium nitrate, car fuel, and a highly unstable substance with a chemical nomenclature of triacetone triperoxide (TATP).

A key challenge in detecting objects such as TATP is that they often pass through typical metal and other object detection platforms currently in use, said Atul Rai, chief executive of artificial intelligence-driven video analytics firm Staqu Technologies.

He said while safety monitoring depends heavily on intelligence reports, select objects that may pose security threats can pass through current systems—even though they are used, monitored, and updated regularly.

What are the key threats that are constantly monitored?

Most public safety threats that are constantly monitored are based on reports from the country's top intelligence agencies. Police forces, including those in Delhi, Haryana, Uttar Pradesh, Punjab, Maharashtra, and other states, deploy some form of image-recognition technology, now including AI, to identify perpetrators on a watch list as part of their regular security monitoring operations.

Law enforcement bodies also proactively monitor aspects such as suspicious loitering and stalling in public spaces, mismatched number plates, or suspicious object shapes, using various safety tools. Most of the current monitoring occurs reactively to intelligence inputs, but proactive monitoring is also picking up.

Can technology make public spaces safer?

Absolutely. Most police forces have long been deploying facial recognition and object recognition tools to enhance safety monitoring, particularly in metropolitan cities. However, unorganized public parking spaces pose a tougher challenge since it is difficult for law enforcement agencies to detect objects inside cars in crowded, open parking lots. The absence of handheld or floor-operated metal detectors or hybrid object scanners, seen often in malls and hotels, makes it more difficult to secure open public parking spaces.

How do tech platforms detect suspicious objects?

Rai’s Staqu, for instance, works with police forces through its safety platform Jarvis, which contributes to 30% of the company’s annual business. These platforms typically tap their own databases as well as publicly available open-source databases of suspicious objects, and then use machine learning models to train algorithms to recognize suspicious objects, their shapes, colours, and other metrics.

Most of these tools are deployed by law enforcement bodies through their own data centres and proprietary platforms, with the technology provider not controlling the data flow and therefore not falling under the ambit of the Digital Personal Data Protection Act, 2023.

What more can be done?

A senior law enforcement officer, on condition of anonymity, added that the country's top investigation agencies continue to work with AI platforms to potentially detect more and more suspicious material through image, object and pattern recognition, live video analysis and more—with a globally-shared list of suspicious objects being constantly updated for safety standards.

However, 100% public safety is difficult to attain, since “safety versus threat actors is a cat-and-mouse game", the officer said.

That's why law enforcement agencies urge proactive citizen reporting for prompt actions through eligible bodies, especially for unorganized public spaces in crowded areas.