Researchers in the USA have developed the first system artificial intelligence that analyses past crime data and can accurately predict 90% where and how many new crimes will occur in a city within the next week. This is the most accurate algorithmic tool developed to date in the field of criminology.
The researchers from the University of Chicago, led by Isanu Chapopadayai, who published the paper in the journal Nature Human Behaviour, said that their «smart» model can, based on historical data, predict with a fair degree of accuracy the probability of specific crimes in individual areas of a large city.
«If you feed the model with data that happened in the past, it will tell you what's going to happen in the future. It's not magic, it has limitations, but we've confirmed that it works really well. You can also use it as a simulation tool to see what will happen if crime increases in one area of the city or if the available police force increases in another area,» Chatopadayai said.
Police authorities and other relevant agencies in different countries are increasingly making use of «smart» predictive technologies. On the other hand, such systems with special forensic algorithms have been criticised for their potential to perpetuate racist and other prejudices. This is denied by the researchers, who argue that their system, which has been tested in eight American cities, will help police to allocate their resources more appropriately and avoid being caught by surprise.
