Xlaw: an algorithm that includes thefts and robberies before they occur. The operation explained by its inventor
There is an algorithm able to predict theft and robbery through an alert system that with circles on a map shows the crimes that are about to occur. Police officers receive the alarm on their mobile or smartphone through an application. The software is called Xlaw and was invented by the inspector Superior Elia Lombardo. The policeman took twenty years of studies and research to make it happen but in the end he managed to create this incredibly useful technological tool for the law enforcement work.
The program has been operating for years in 9 Italian police stations including Naples, Salerno, Venice, Livorno, Modena, Trieste, Trento and its utility has been certified from one study of Professor Giacomo di Gennaro of the University of the studies of Naples Federico II and of Professor Riccardo Marselli of the Parthenope of Naples.
The inspector-scientist told his story in a book published for Mazzanti Libri, in which he explains how the intuition of the algorithm was born. It is a discovery that according to Lombardo could completely change the way to fight crime: no longer a method based on the pursuit of the thief to the crime occurred, but a strategy based on the prevention of crime.
According to the first statistics in Naples the software has a precision between 87% and the 93, in Venice between 92 – 93 % and in Prato around 94%.
The mathematical model of the algorithm based on the “guard and thieves game” to predict thefts and robberies
The magazine Business Insider Italy interviewed the inspector who explained in detail the operation of the algorithm.
Lombardo explains that the traditional system of territorial control is also based on statistics and forecasts. In fact, “action plans” are normally established on which to place patrols.
This approach, however, according to the inspector, offers few results and it is much more effective to be able to know exactly where a crime is about to occur: “I wanted to change totally paradigm and really do prevention. The results then came: first we had street patrols waiting for someone to commit a crime. Today with XLAW we know exactly where to look: patrols go down the street to prevent something from happening. It seems nonsense, but it is a revolution “.
The system of the guard and the thief would be based on a mathematical model where the leader is the thief. The criminal in fact has more information than the policeman who in this dynamic represents only the pursuer or “follower”
The new strategy of the algorithm to predict theft and robbery: the reversal of roles between thief and policeman
“He who commits crime is a leader in this game because he has more information. He knows the prey, he knows the enemy, he knows the working hours of the policemen, he is aware that the police are conducting methodical operations, so he knows well at what time and where the police carry out routine checks. But this process had to be reversed. “
To change the rules of the “game” and transform the policeman from follower to leader it was necessary to make sure that the police had an “advantage” over the thief in terms of information and data.
“I took 20 years of data and crimes and I studied them one by one, so I didn't stop at the theory. I read the complaints, I went out into the street, even at night, to see the places where the crimes took place. After years of study I discovered that among the many crimes that human beings can commit, many of them are predictable. These are the so-called predatory crimes: theft, robbery, pickpocketing. These crimes are predictable because the people who commit them are organized, in short, they are professionals. Anyone who carries out a professional activity, which in sociology and psychology is called routine activity, tends to choose a strategy “.
The criteria according to which the thief acts and the prediction of the crime areas
The trick is to understand the mechanism behind the strategy to unhinge it from within. According to Lombardo there are two criteria on the basis of which the criminal acts: the subjective and the objective. Those who follow the objective principle choose a precise place where it finds the presence of preys and targets
Contrary to what is thought, it is therefore more likely that crime will be concentrated in the “rich” areas of the city rather than in the poorer and degraded areas: “In the past the theories asserted that crime occurs where there is more degradation, while currently it is believed that predatory crime occurs instead where there is more wealth, where there are more prey “, declares Lombardo.
The criminal who follows the subjective criterion chooses instead to act where he feels safer and has more escape routes: “Collecting so much data I understood that the crime predatory creates real hunting reserves ”. Hunting reserves are defined on the basis of wealth but also based on the presence of temporal “regularities”: entry and exit from homes, offices, schools, markets, commercial establishments, arrival of trains, buses, cruise ships.
The virtual city software
“I created two layers within the software: in the first I created a sort of virtual city, where I entered all the socio-demographic information of the city. Number of citizens, number of homes, shops, banks, post offices, train timetables, bus timetables, cruise ship schedules, in short, I have included all the information needed to understand how the city operates and moves “, he points out.
The soldier then tried to define the psychology and characteristics of the thief: “the type of author (young, old, foreign), type of day, time , the type of prey (student, professional, senior, housewife) and type of target (clock, money, iPhone, vehicle). This information defines a criminal model “.
Cross-referencing all these data leads to the identification of the crime: “In the series of phases to regular operations within the city (the day of payment of the pension, the pool where the mother accompanies her son to 17: 00) and in particular in game reserves, if we superimpose the criminal data on this virtual city, we realize that the crimes take place in those game reserves and correspond precisely with the phases to regular operations within the city ”, he concludes.
Patrolling the “game reserves” removes elements of regularity and predictability to the criminal and makes it more insecure. The idea is that if you are in the right place at the moment the thief will end up giving up theft alone
The comparison with medicine
However, the science of investigation and prevention is not exact and mathematically definable at 100, is rather “a heuristic science like medicine” explains Lombardo. It is based on signs and symptoms and can only make probable predictions.
Through a formula Xlaw can also measure the risk index of a city and the data are surprising. “I have the criminal pressure index on a monthly basis: in Naples it is 2, and it is low, in Venice it is 6 and Modena is 11. Clearly this number is related to the number of events, population, etc. This is also a paradigm shift: demonstrating with data that one thing is the perception of risk, reality is another. “
The evolution of the algorithm
Xlaw is an artificial intelligence that varies over time based on the data of the complaints: “For example, the type of crime 'drug dealing in the home' did not exist before . I developed it for predatory crimes, but the system is beginning to create a series of criminal models that I had not foreseen. For example, in the 2004 I did not have the baby gangs, now it is starting to configure these too criminal models “. The system states Lomabrdo is evolving. The hope of the police is that in the future it can be extended to all territories.
Born in the province of Pisa on 26 June 1991. Graduated in Philosophy and graduated from the School of Journalism of the Basso Foundation of Rome. Together with two colleagues he won the Morrione Prize 2018 with an international investigation into the weapons factory RWM in Sardinia. He collaborates with TPI and with the Manifesto, Il Fatto Quotidiano and with the Report program