By Dr. Anand Kulkarni, Professor and Associate Director, Institute of Artificial Intelligence, MIT World Peace University, Pune
Prevention is better than cure; we all know this and have certainly experienced it. Just as the chances of accidents can be reduced with proper care, measures such as awareness and education among people are currently being used to reduce crime. Similarly, post-crime measures include education, punishment, surveillance, amendment of existing laws, creation of new laws. Violence, theft, and crimes related to women are increasing every year worldwide. With the ever-growing population, it is becoming increasingly important to be proactive in preventing all types of crimes. This is vital for maintaining social health. In this context, there is no alternative but to turn to advanced technology. The use of artificial intelligence (AI) can be of great help to the police and intelligence in effectively preventing crime by providing advance notice or alerts. This can aid in detecting the possibility of crime and taking preventive measures and proactive steps.
We know that the probability and amount of rainfall, direction of storms, etc. can be predicted from the signs of climate change and appropriate measures can be taken. Machine Learning method plays a crucial role in identifying patterns in past data and making precise future predictions. Recently, ML methods have also shown strong potential in predicting the possibility of crime. According to expert studies, in general, if a criminal is successful, the mode, nature, time, and place of the crime do not change; they become consistent in committing crimes in a certain way, at a certain time, and in a certain place. Just as criminals learn, ML algorithms also learn the patterns of past crimes and predict the time, place and likelihood of the next crime. This helps the concerned departments to become active.
Importantly, as crimes are committed or solved, the algorithms and software based on them become more sophisticated. Interestingly, these algorithms use information obtained from various sources without any human intervention, reducing the chances of bias or error.
With the increasing use of mobile phones and the internet, information is flooding in the form of photos, videos, audios, emails, letters, newspaper reports, speeches, social media posts, locations, and more. Various AI & ML-based algorithms are used to quickly analyse this data, perform mathematical transformations, draw inferences, and reach conclusions. This makes it possible to predict the likelihood of crimes in a very short time, which is almost impossible otherwise. A smart city-related company in Israel called ‘Cortica’ has developed software that analyzes the information obtained through CCTV. This software utilizes certain AI algorithms to recognize the faces in a crowd, identify crowd behavior and movement, and predict the likelihood and nature of a crime. Interestingly, these intelligent algorithms make it possible to analyze several terabytes of video footage in minimal time and make quite precise inferences.
An intelligent software has been developed by Cardiff University in England. It is being used by the US Department of Justice. It employs several ML algorithms classifying the posts on social media platforms for predicting the nature, time, location, etc. of a possible crime by an individual or a mob. ‘Predpol’ is another intelligent software tool that categorizes the city of Los Angeles into smaller segments based on crime information. This has enabled the police to focus on specific areas, leading to a 30% drop in theft, and a 21% reduction in violent crime. Nowadays, smart AI & ML-based intelligent surveillance cameras can identify the person at the door. It informs the owners if the person is safe to let in, and notify the police otherwise.
The safety of children and women is an important issue in India, as well as in almost every country in the world. Recent incidents have only added to these concerns. Some mobile ML-based apps like ‘Safetipin’ and ‘Savdhan’ are currently available in the market. Based on people’s feedback, good and bad experiences in the city, rush hours, etc., these apps suggest alternative safe ways for women to travel, suggest a place in the city, time, etc. to avoid.
The app on the mobile phone can be activated with one click as soon as there is an indication of some untoward incident. Immediately it starts recording the surrounding sounds, places, etc. It transmits all the recordings to the parents or next of kin, even calls them. Importantly, these apps can be used for security systems to identify unsafe and less secure locations in the city. The Department of Homeland Security in the United States employs ML-based algorithms to precisely identify the race of the perpetrator based on the characteristics and structure of the DNA obtained in the case of violence against women. This is proven to be very useful in a multi-ethnic country like USA as well as in India. Due to this, it is possible to solve many serious and complex crimes against women.
It is important to mention that the AI & ML based algorithms provide advice or make predictions using experiences, previous cases, facts, etc. Therefore, they can be valuable guides for executives and officers. The prevention of crime, early detection, and assistance in decision-making can increase safety as well as save valuable time of the courts. It is also possible to set up an institute for crime prevention using AI on the lines of Forensic Science University in Gujarat. There will always be a need for such organizations for nation building. There are endless opportunities for research in this. This may not only give an edge to the police and intelligence over the criminals, but also our researchers can make a valuable contribution to safety and well-being of society.