AI predicts crime a week in advance with 90 per cent accuracy

It is possible to predict the location and rate of crime a week in advance with up to 90 percent accuracy. The same systems have been shown to perpetuate racist bias in policing, but the researchers who created this artificial intelligence claim that it can also be used to expose those biases.

Ishanu Chattopadhyay and his colleagues at the University of Chicago created an artificial intelligence model that analysed historical crime data from Chicago, Illinois, from the end of 2016 to 2014–2018.

The model was able to predict the likelihood of certain crimes a week in advance, with up to 90 percent accuracy. It was tested and trained on data for seven other major US cities.

Past attempts to use artificial intelligence to predict crime have been controversial. In the last few years, the Chicago Police Department has been experimenting with a list of people who are most likely to be involved in a shooting. When the list was finally released, it was found that 56 percent of Black men in the city were on it.

Efforts have been made to reduce the effect of bias and the artificial intelligence doesn't identify suspects, only potential sites of crime. He says it's not a report about minorities.

There are law enforcement resources. You would like to use that in an optimal way. He wants to know where homicides are going to occur.

Read more: Google wants to challenge AI with 200 tasks to replace the Turing test

Rather than being used to allocate police resources, the predictions of the artificial intelligence could be used to inform policy at a higher level. The data used in the study has been made public so that other researchers can look at it.

The researchers looked for areas where bias is affecting policing. They looked at the number of arrests in neighborhoods with different levels of poverty. It was shown that crimes in wealthier areas resulted in more arrests than they did in poorer areas.

Lawrence Sherman at the Cambridge Centre for Evidence-Based Policing, UK, is concerned about the inclusion of reactive and proactive policing data in the study, or crimes that tend to be recorded because people report them and crimes that tend to be recorded because police go out and look for them. He says that the latter type of data is vulnerable to bias. He says it could be a reflection of police discrimination.

Nature Human Behavior was published in the journal.