Today's world is intrinsically or extrinsically connected to the technology at the hip. Despite age or gender barriers, society is evolving along with technological advancements in a manner never imagined before.  

In the third edition of the NASSCOM Xperience AI virtual summit, industry experts shared the use cases of how AI is being deployed to crack criminal challenges, public safety and has helped ease public infrastructure. 

The virtual discussion was moderated by Gopali Raval Contractor, Managing Director, Cloud First, Advanced Technology Centers in India, Accenture. Atul Rai, Co-founder & CEO of Staqu technologies and Ranjeet Deshmukh, founder of Roadbounce, shared their insights on their AI applications and use cases. 

Enhancing road safety with AI 

Ranjeet Deshmukh was already the founder of two successful startups while thinking about a problem statement for his third venture. His interaction with government officials gave birth to the idea of Roadbounce. Roadbounce is now digitizing the process of finding damaged roads in major cities of India and Singapore.  

"India is the second largest road network in the world. So how can the authorities know where the potholes are after every monsoon" asks Ranjeet.  

With the Government of India, they are working on means to enhance road safety and bring down traffic accidents due to bad road conditions. 

Fighting crimes with AI 

In a nation like India, where statistics prove that safety from crime is a major concern, Atul Rai's Staqu technologies are a promise for a secure tomorrow. "Every hour, we are having 300 IPC crimes and 100 murders in India", says Atul. 

70% of criminals in India are repetitive since individuals bailed from jail are committing the crime continually. CCTV cameras are in every nook and corner of our streets, but their efficiency is zero without AI input.  

To keep suspicious individuals under the police radar, Staqu technology's Jarvis functions with homeland security and intelligence. As a result, over 45 lacks people with criminal records are being monitored in real-time. 

Improving accuracy and fighting challenges 

Jarvis only carries the information of individuals with a criminal background. Along with visual data, voice recognition is also enabled in the model, which improves the accuracy of identification. Jarvis has successfully recognized 10,000+ criminals and eight terrorist activities in the country. 

Roadbounce makes use of vibration in vehicles to identify potholes and data collection. The accuracy rate of visual AI might be limited in the model if the pothole is flooded with water or covered by another vehicle.  

Developing a centralized database for Roadbounce was time-consuming as the vibration pattern of every vehicle varies. Rather than depending on the civilians who might choose to drive through better roads than the ones with potholes, the data for the model is collected by the team. 

Building business and lack of experience working with the Government was yet another challenge that the company initially faced but gradually overcame. "Now the business we have is either from the government or through government influence", says Ranjeet.  

Where is AI headed? 

AI has evolved from concentrating on one aspect of data. It now combines all kinds of data sources such as audio, video and NLP. Atul Rai believes that this is the aspect that will make AI more powerful.  

The world is moving ahead as we always thought it would. Ranjeet opines that "AI or any other technology is just a tool. When it comes to destroying humanity, it would be humans themselves".