Spanish police now use a system of artificial intelligence that can help to recognize false reports of theft and other crimes. Researchers from Cardiff University and Charles III University of Madrid developed the AI system, which was baptized VeriPol.
The software uses automatic text analysis and machine learning to identify false claims. According to computer scientists, VeriPol is able to correctly recognize false messages in more than 80 percent of the cases. This could potentially relieve the police of a great deal of pressure.
Submitting a false declaration is punishable in many countries. However, when it comes to theft, for example, a tax return can help you to file an insurance claim, for example. As the police in many countries are understaffed, departments often have little opportunity to investigate false reports. That’s what makes declarations slip through the net.
However, VeriPol is able to analyse written statements. This allows patterns to be recognized that are often associated with false statements. Think of the type of object that would have been stolen, the descriptions given of robbers and other indications.
Themes that often seem to indicate a false report, focus on the stolen objects and not on the incident. A lack of detail indicates that this is an invented event. The fact that no 112 call is made immediately after a robbery is often also an indication that the declaration is false.
More efficient use of resources
VeriPol uses natural language processing to detect these false declarations. This is an element of machine learning that helps artificial intelligence to understand the natural use of human language. Algorithms are used to decode that language and VeriPol uses previous returns to strengthen the system.
The researchers believe that the AI tool can help in the choice of where to put certain resources. If a report is considered to be potentially false, people could investigate it further. At the same time, the software could also discourage people from making false declarations.This news article was automatically translated from Dutch to give Techzine.eu a head start. All news articles after September 1, 2019 are written in native English and NOT translated. All our background stories are written in native English as well. For more information read our launch article.