Crime prediction techniques have often been under scrutiny because they lack objectivity by virtue of ingrained biases

Artificial intelligence is gradually catching up with advanced policing, pricking the interests of Governments. Who wouldn't want to have policing tools in their armor for the rate of crime is only seeing an uptick with every passing day? Crime prediction techniques have often been under scrutiny because they lack objectivity by virtue of ingrained biases. Given the facets and subjectivity surrounding the crime, the predictive techniques couldn't gain as much support to become a mainstream technique. A new crime prediction algorithm developed by data scientists from the University of Chicago forecasts crime and could predict violent crimes a week in advance with 90% accuracy. The AI tool was tested for the Chicago city area along with a few other major US cities and found similar results. The AI tool could predict crime by observing the patterns of data on violent and property crimes from a particular geographic location.

 The research team, in another model, also studied the police response to such crimes from locations with different socio-economic statuses. They found the number of arrests in the wealthier region was more when compared to socially and economically backward areas. The bias in police attitude was apparent as the crime in poor localities didn't result in many arrests.

In earlier attempts of predicting crime, Chicago police couldn't find much success. They followed a similar predictive algorithm only to find it biased. Its task was to list people who are most at risk of being involved in shooting as perpetrators or victims. But when the list was finally released, it was found that nearly 56% of the Black men in the city were part of the list. In the opinion of Ishanu Chattopadhyay, Assistant Professor of Medicine at UChicago, and senior author of the new study, it is a direct consequence of putting the system under too much pressure. More resources are spent on making more arrests in wealthier areas making underprivileged locations deprived of resources. Chattopadhyaya says prediction tools can be employed to help frame policy at higher levels than helping police in resource allotment.

Historical data from the City of Chicago around two categories – violent crimes and property crimes – was used to test and validate the AI tool. The used data is considered most valid because the areas had the utmost likelihood of crime occurrence. Historically, the areas selected were under police observation for lack of distrust and cooperation. These locations are less likely to yield to enforcement bias unlike in cases of drug crimes, traffic crimes, etc. Crime prediction has been applied earlier too but the difference lies in the approach. Those approaches merely focused on crime hotspots, completely ignoring the complex social environments of the cities and the complex relationship crime shares with police enforcement. As per a statement by James Evans sociologist and co-author, this tool can enable the discovery of socio-cultural connections too.

The tool works by isolating crime by observing spatial and temporal data of discrete events to detect patterns. Instead of relying on conventional boundaries, it divides the region into spatial units, to be able to eliminate the element of bias. Evans says, by doing do, it will let us see the crime prediction from an entirely different point of view, ask new questions, and evaluate police action accordingly. While Chattopadhyay is very optimistic about the tool as a means of learning for other researchers, he doesn't shy away from acknowledging its susceptibility to biases. "It's a minority report", he says.

"We created a digital twin of urban environments. If you feed it data from what happened in the past, it will tell you what's going to happen in the future. It's not magical, there are limitations, but we validated it and it works well," Chattopadhyay said. "Now you can use this as a simulation tool to see what happens if crime goes up in one area of the city, or there is increased enforcement in another area. If you apply all these different variables, you can see how the systems evolve in response."

More Trending Stories 

• Ethereum Rally is Possible Only if ETH Crosses US$1,500 Resistance
• Google Cloud Paves the Way for Users to Track Down their Carbon Footprint
• Meta's Metaverse is Not a Place for LGBTQ+ Community, Just like Insta
• This Robot Is Prejudiced and Creators are Benefiting from its ML Model
• Can You Land a Job in FAANG as a Self-Taught Cybersecurity Pro?
• Researchers Infuse 'Human Guesses' in Robots to Navigate Blind Spots
• Quantum Internet Seems Possible! Qubits can be Read Without Being Destroyed