Artificial intelligence catches criminals
The Russian company NTechLab is engaged in developments in the field of
NTechLab's main client is law enforcementorgans. In the spring of 2019, with the help of FindFace, the police detained 11 criminals in Tatarstan: the system matched the faces caught in the lens with the database of the criminals' pictures. At the FIFA World Cup in Russia, FindFace helped to catch more than 180 violators - including the sponsorship cup thief.
How to recognize the face of an offender?
Chinese startup develops technologyrecognition of faces, voice and fingerprints, as well as analyzes the movements and behavior of people - the police can use it to predict and prevent possible crimes.
As of 2017, the company helped solve more than 10 thousand crimes.
Can you trust a lie detector?
Today, there are many options for expertise that allows minimizing the human factor.For example, the analysis of traces of DNA at the crime scene, but it is not always possible to find them.
Once upon a time, criminologists pinned their hopes onpolygraph. It records a complex picture of body reactions, such as changes in skin electrical conductivity, pressure, and heartbeat. In theory, if a person is lying, the body will give him away. In practice, it has not yet been possible to find evidence that there is a pattern of physiological reactions unique to deception.
A sincere person can seem like a liar if he is very nervous, and a rogue one can be calm like a tank, giving out a thoughtful legend.
What can brain activity say?
In the 1960s, researchers Robert Chapman andHenry Bragdon studied patients with neurological disorders and made a discovery. They found that the brain responds differently to meaningful and insignificant visual stimuli (images). And most importantly, this reaction can be tracked using EEG sensors.
If the image is significant (for example, a familiar face),approximately 300 milliseconds after its demonstration, a deviation appears on the tape - a «signal». Based on this discovery, today, for example, interfaces for «mental» commands to the computer.
Many years later, other scientists - neuroscientistsLawrence Farwell and his teacher Emmanuel Donchin from the University of Illinois suggested that this technology could be adapted for the needs of police and intelligence agencies.
Farwell didn't give up. He conducted an experiment in which FBI employees and volunteers from a control group were presented with words on a screen that only people from the intelligence agencies could know.
The results were impressive: in 100% of cases, the machine correctly indicated which of the participants knew the words and who did not. In addition, he managed to achieve positive court decisions - albeit indirectly.
How to accurately calculate the crime scene?
The company uses AI to predictthe alleged place, time and type of crime. PredPol technology was originally developed by the Los Angeles Police Department and UCLA University and is now being developed by Doctors of Mathematics, Forensic Science, and Sociology in collaboration with police officers.
An algorithm developed by PredPol analyzeseconomic situation, social trends and human behavior, and compares information with crime data. It calculates the areas (with an accuracy of 50 m²) and time intervals in which a crime is most likely to occur.
The police can choose the type of crime -for example, robbery, hooliganism or vandalism. The program generates a map based on Google Maps showing where the selected types of crimes are likely to occur.
Security robots
Knightscope security robots help the US police patrol potentially dangerous areas: parking lots, large shopping malls, sports stadium lobbies.
They are able to recognize faces, read license plates and create a thermal image. Thanks to this, Knightscope robots can distinguish an ordinary passer-by from a suspicious person.
Finding criminals on the darknet
In August 2015, Canadian law enforcement officers ordered the development of a search robot that explores the depths of the Deep Web. He is looking for offers of illegal goods and services.
The software part was prepared by Mercur ITSolutions, which has collaborated with police before. Also in Canada, on August 30, 2016, a woman was detained for purchasing a deadly radioactive element, polonium-210, over the Internet.
Police department in an American cityBoston began a new program in early January 2017 that collates data from the Deep Web and social networks. According to the plan, development of the program will cost $1.4 million.
Commissioner Evans said in an interview with Bostonradio that it is “an essential tool of the rule of law that will help keep our neighborhoods free from violence and terrorism, prevent human trafficking and protect children from pedophiles.”
Documents obtained by the Boston Globe showed thatthe software will show the police the geolocation of possible offenses in real time. Previously, US police have already successfully closed Silk Road, Silk Road 2, Black Market Reloaded and many other underground markets.
How are Russian security forces using technology?
The Ministry of Internal Affairs plans to use neural networks forinvestigating serial crimes and drawing up a "composite" based on the violator's DNA. Technically, such a possibility already exists, but in Russia there is still insufficient genomic data base for analysis.
The ministry intends to actively use artificial intelligence in its work, namely:
- in software, which should automatically detect signs of serial (interconnected) crimes;
- in software that will allowdetermine the external anatomical signs of criminals (eye and hair color, face and head shape) based on biomaterial obtained from crime scenes, for example, blood traces.
Read more:
Physicists have created an analogue of a black hole and confirmed Hawking's theory. Where it leads?
Sea slugs cut off their heads to regenerate a new body
Abortion and science: what will happen to the children who will give birth