Artificial intelligence and antiterrorist operations

The advantages of AI technology

The terrorists attacks executed in some democratic states are trying to create panic and fear among the population and undermine the public support for the authorities of the targeted states. Even when the consequences of a terrorist attack are effectively managed, such events should better be prevented. Therefore, most of the states are focused on preventing terrorism, as part of the national security strategy.

There are two ways when it comes to preventing terrorist attacks. The first method is deterrence, which consists in protecting the critical infrastructure, applying the security controls and consolidating the punishments through legislative acts. The second method is limiting the operational capacity of individuals, by arresting terrorists before their attacks are actually executed, countering the recruitment and radicalization and imposing movement restrictions or the freedom of circulation (declaring them undesirable people and expelling them to their origin country).

Both methods have common features. Precisely, they ask for important human and financial resources and, for some of the cases, they are breaching the rights of those who are applied for, when the terrorist can rely on procedures allowed by the principles of democracy and can get away with the punishment or even ask for indemnification for the moral damages caused by authorities’ action. In order to get to that balance between preventing attacks and following the citizens’ life, the two methods mentioned above should be improved and this can only happen through prediction.

To that end, technology based on artificial intelligence might allow the collection of data and metadata from different relevant digital systems and may offer possible well-constructed developments to the institutions.  Therefore, the use of AI in counterterrorism is focused on making predictions, which are making the use of available data more effective, are minimizing the intrusion in citizens’ private life and are diminishing the biases in making decisions. In an era of social networks and IT, when it is easier and easier to collect, archive and analyze the activity of people and companies, the predictive methods used by the AI technology may prove to be important, given that these can direct the resources of authorities towards areas, individuals and legal persons who can represent threats, concurrently with reducing the number of people subjected to more invasive surveillance.

Possible methods to use the AI technology

Recent developments in AI technology make possible the analysis of individuals’ activity to predict certain attacks or to identify possible terrorists based on the activity of a group. The necessary information for this analysis are collected from the online communication metadata, the login in different internet connections, online shopping, online reservations and posts in a social media group or even data provided by the GPS system included in the current digital device.  Based on the analysis of these data, they could predict:

The date and place of the execution of an attack

Currently, they are making efforts towards developing models that could predict the place and time of some attacks. This development direction of the AI technology is based on estimations on the possibility of an event after a similar one, something observed also in the case of crimes like robbery.

Also, another line of development when it comes to predicting the place and time of an attack was based on foreign factors – like the political context – over the incidence of terrorist attacks. Until now, this line of development did not give the expected results.

The vulnerability to radicalization and recruitment

For the terrorist organizations and the extremists, the internet is a tool wherein they are executing the psychological warfare, spreading propaganda and disinformation and attracting new recruits. Therefore, the social networks have become their uppermost used method to recruit vulnerable people.

The lack of proper tools for the effective identification of online messages having content related to terrorism can lead to ignoring a recruitment activity or even the planning of an attack, having dramatic consequences for the citizens of the state targeted by the attackers. 

As the companies offering online services commit to more and more responsibilities related to protecting their users, they have developed tools allowing the identification of a suicidal risk or vulnerabilities to mental illness, tools that could also be used to identify individuals who might adhere to violent extremist ideologies.

At the same time, the terrorist threat on the European territory has pushed the European Council to provoke the European industry to develop new technologies and methods to automatically detect and eliminate the context promoting terrorist and terrorist acts. Consequently, it was initiated the “RED-Alert” project (An early warning system for online terrorism content, based on the processing the natural language, analyzing the social networks, artificial intelligence and the complex processing of events), funded by EU. Project’s purpose is the development of technologies able to detect the online radicalization and support global antiterrorism efforts. The project is coordinated by the SIVECO company from Romania. The project, started on June 1^st 2017, was completed on May 30^th 2020 and it was budgeted with 5 million euro. The results of the research will be available for the public in the following period.

Terrorists’ identification

In 2017, the National Security Agency (NSA) used SkyNet within the operations from Pakistan. Within the application have been used the metadata provided by the phones of 55 million Pakistanis. Although the used model did not give the expected results, the use of such an application shows how metadata can predict activities related to terrorism.

Another identification method for terrorists is facial recognition. Today, the local police and the American rule of law agencies have a biggest chance to anticipate and prevent terrorist actions through the new artificial intelligence methods.

Facial recognition allows the authorities to clearly identify any person, to compare billions of public images collected from open sources and identify a certain person. The application, developed by the company Clearview AI is currently successfully used by thousands of employees of the US authorities to combat crimes.

Conclusions

The still small numbers of uses of AI technologies in preventing and combating terrorism, as well as the results they got so far are rather suggesting possible ways to use this technology than powerful proofs of reaching a certain operational level in the field. Therefore, we must be aware that the immediate transformations related to the use of AI in preventing and combating terrorism cannot have the expected results but, at the same time, nor should they abandon the research in the field.

AI can currently be used to make predictions on terrorism based on metadata, intelligence on financial transactions, travel patterns and internet activity, public information from the social networks. Based on that, the authorities can continue their data analysis and interfere when necessary. Supposing that the digitalization tendency continues and that AI technology’s performances will improve, we expect to have also a positive impulse for the use of technologies in preventing terrorist acts.

Translated by Andreea Soare