For what seems like forever, governments and militaries around the world have tried to tackle terrorism with billions of dollars of investment and several conflicts. As technology has advanced over the years, the prospect of using machine learning against terrorism promises to bring the world closer to being terrorism-free.
Terrorist attacks of any scale are never isolated or spontaneous events. If an attack has occurred today, then it almost certainly has been a while in the making, with several intelligent and strategic individuals working together, and several pieces—intelligence failure, lack of preparation of security forces, crowded target location, amongst others—falling into place to ensure perfect execution. Take the 9/11 attacks, for example. Investigations in the aftermath of the tragedy would reveal that its seeds were apparently sown when Khalid Sheikh Mohammed and Osama bin Laden, the two masterminds behind the devastating quartet of plane hijackings, first met in Tora Bora, Afghanistan, in 1996. Likewise, Abdelhamid Abaaoud, the “ringleader” of the 2015 Paris terror attacks, allegedly commenced planning the event when he joined the militant group ISIS in 2014. Although law enforcement bodies and anti-terror forces around the world cannot prevent indoctrination and radicalization of the most vulnerable—the meticulous grooming process through which disillusioned individuals are driven into executing terror operations—they can use a formidable resource to predict and prevent prospective terror attacks – big data.
At this point, the power of big data is well known. After all, elections have literally been won by politicians leveraging big data competently for the purpose. So, by using continuously evolving data related to prime suspects, weaponry procurement and other factors that may precede terrorist attacks via technologies such as computer vision, blockchain, NLP and machine learning, countries around the world can take measures to prevent or mitigate attacks. Employing machine learning against terrorism, especially, can be a golden idea for anti-terror squads with the resources to do so.
Predicting Terror Attacks with AI
Terrorist activities are, more often than not, acts of retaliation by individuals or community-based groups who feel that they’ve been wronged. Investigations after an attack generally tend to reveal that such people felt helpless, afraid, angry, aggressive and possessed intolerance towards those who supposedly wronged them. People going through such a maelstrom of negative emotions are more likely to participate in a hostile public attack than others. Big data and sentiment analytics allow public agencies to gauge, up to a certain degree, the emotional state of citizens and pre-emptively resolve any issues. Sentiment analytics, a futuristic smart city concept for now, involves the in-depth evaluation of text and media shared online by the general public. The sources for this ever-growing data can be social media pages, emails to the government, online forums, social media comments and other digital places where people can air their opinions. Sentiment analysis involves neural networks assessing patterns within such massive swathes of data to predict a probable attack in the future. This is one of the ways in which law enforcement agencies can ensure effective counterterrorism and prevent incidents like school shootings at the hands of disgruntled, degenerate individuals.
In 2021, a team of researchers at the Zhejiang University, Zhejiang, China, worked on a model that used machine learning against terrorism driven by politics or sociological factors. The Dr. Andre Python-led research used AI and machine learning to sift through terrorism attacks that occurred between 2000 and 2006 across 13 undisclosed locations. Over these years, models of machine learning against terrorism were trained and developed to identify six specific variables in the selected locations. In the research, these variables were pre-defined to be key indicators of a terrorist attack.
As stated earlier, terrorist attacks involve detailed planning, training and resource management. Although terrorists will take various measures to ensure that there is no paper trail related to an attack, data, in one form or another, will inadvertently be generated. When public bodies employ machine learning against terrorism, this data can be analyzed and used to forecast the location and scale of an attack. In this way, big data provides the fuel for technologies such as computer vision and machine learning against terrorism.
Conducting Mass Surveillance with Computer Vision
Terrorist attacks generally occur in high-density, highly crowded public spaces to maximize the impact. Therefore, places like crowded subways, railway stations, shopping malls and tourist zones need machine learning and computer vision-guided autonomous, continuous surveillance. In smart cities, such systems can be trained with datasets containing information about suspected individuals’ social security details—name, the content of text messages and emails, year of birth, immigration status, amongst others—to classify them as security threats. Intelligence agencies can then closely monitor such persons in a crowded spot.
In smart cities, computer vision-powered CCTV cameras can run facial recognition of thousands of people at once. Computer vision algorithms serve two purposes – firstly, suspected criminals and terrorists can be picked out from a large and moving crowd, and secondly, the feature allows law enforcement officials to monitor the movement of such persons of interest. This will help the officials to apprehend the suspect and the people who work with them on future attacks. Apart from identifying people and crowd management, AI-powered tools can also detect whether someone within a crowd is carrying a gun and other weapons.
The successful implementation of computer vision and machine learning against terrorism-based activities depends upon the generation of continuous, diverse and accurate data. Therefore, smart cities are arguably the most likely places to implement machine learning against terrorism as they have several IoT-based data receptors at multiple locations.
Tracking Terror Financing with AI
Cutting off the financial supply network of terrorist groups is an effective way to reduce the number of attacks any country encounters annually. However, tracking such payments is a tall task, especially as terrorist collectives not only use cutting-edge technology but also employ seasoned professionals for executing various disruptive functions—hacking into a company’s IT network, duplicating ID cards online, manipulating traffic lights remotely and many others. Normally, terrorists use inconspicuous cyber-threats to stealthily move funds without detection.
As you know, machine learning is all about evaluating massive amounts of changing data and finding patterns. Machine learning-based tools track the behavior of an account holder to predict suspicious activity. Certain actions, such as abrupt changes in the buying patterns of a specific individual, frequency of cash withdrawals, regions in which cash is withdrawn from ATMs and unusual dormancy in purchases are treated as red flags. Machine learning algorithms rely on big data generated at banks, ATMs and various stores.
With financial behavioral monitoring, anti-terror agencies can keep an eye on the places from where weapons are procured. Weapon procurement is normally the phase that comes after radicalization and strategy formulation, making it, arguably, the stage when the attack planners are at their most vulnerable. In other words, this is the phase when governments can exercise maximum efforts to thwart the terror problem just as it takes shape. In this way, countries can bust terror operations at the very beginning, simply by using the vast reservoirs of data that can be collected from various sources.
Detecting Online Terror Activity with Machine Learning
Terrorist collectives have adapted to digitization by carrying out attacks or spreading hate propaganda within the digital realm. Preventing such kind of cyberterrorism is possible with the help of big data analytics. Essentially, anti-terror bodies can leverage machine learning and NLP to find dangerous content on various forums, websites, social media pages and other online places. Already, the emergence of AI in cybersecurity allows businesses and governments to deal with all kinds of data security threats. Similarly, AI can also act as a shining light against online terrorism activities and hate-fueled content. Social media companies need to have regular checks to remove propaganda videos or photos on their platforms that may contribute towards the radicalization of wayward youngsters.
The application of big data and machine learning against terrorism gives anti-terror and law enforcement agencies a much-needed headstart against future attacks. To prevent macabre attacks like the ones witnessed this side of the millennium in New York, Paris, Christchurch or Mumbai, such a headstart is invaluable for citizens and governments. Having the right kind of hardware and software infrastructure and, most importantly, the data to underpin all operations makes living without fear in today’s turbulent times slightly easier for everybody.
