Preventing crime before it occurs? Dubai Police has signed an MoU with Indian startup, Staqu Technologies, for its Artificial Intelligence (AI)-based solution on ‘predictive policing’.

This concept revolves around utilising AI tools for curbing crime before it occurs through practices that can analyse huge amounts of data from different sources in real-time and generate decision metrics. The Dubai Police aims to curb violent crime by 25 per cent by 2021.

Founded in 2016, Staqu had earlier raised $500,000 from Indian investors and hopes to raise another $3 million to $4million this year in a Series A funding round. It was hand-picked by Dubai Police, apart from another US-based start-up, from amongst 677 applicants from across the world, Atul Rai, co-founder and CEO, told BusinessLine .

The startup is now looking to expand in countries in West Asia, , Africa and the South Asian regions.

The Gurugram-based firm’s Artificial Intelligence-Based Human Efface Detection (ABHED) enables police to digitise criminal and missing persons’ records and retrieve this information in real-time. This radical disruption helps the security forces track down suspects and prevent crime before it occurs.

Following a successful pilot programme with the Alwar Police in Rajasthan, ABHED has been adopted by police in Punjab and Uttarakhand as well. Besides, Staqu is also helping the National Crime Records Bureau (NCRB) and the State CRBs.

The solutions are packaged in superior AI software that can be integrated with Smart Glasses, empowering security forces to scan a crowd through image search and look for any miscreants, terrorists or criminals.

Staqu’s innovation has led to the creation of various systems. Its Smart Glass technology enables police to get real-time information on anti-social elements and empowers them to analyse a crowd and segregate individuals based on their behaviour patterns, resulting in more precise actions.

In India, police usually have raw and incoherent data from diverse sources, which is the foremost problem faced by any predictive policing activity. Activities of the police in real-time is fraught with challenges such as identifying criminals or predicting criminal activities based on such data and the lack of adequate technology that can extract and connect in real-time, he said.

This data is often categorised as speech, image and pre-recorded non-uniform textual data about crime and criminals. Simple software algorithms are unable to help compute such complex data structures. It is here that AI assists. For instance, to spot an individual and his/ her chances of being associated with a crime can be predicted through a specific set of hybrid AI technology, which amalgamates distinct neural network models to process image, speech and text, to extract meaningful information in real-time and reconcile it with the database, aiding in concrete decision-making.

Through the amalgamation of these data points in real time, the police can effectively predict parameters related to crime with a higher accuracy rate, which has been difficult to achieve with existing intelligence measures. Also, having an AI-based database would solve problems arising from the frequent transfer of police officials, who have to start afresh each time they go to a new place.