Stories
Slash Boxes
Comments

SoylentNews is people

posted by janrinok on Saturday July 09 2022, @10:25PM   Printer-friendly
from the patterns-of-life dept.

A new computer model uses publicly available data to predict crime accurately in eight U.S. cities:

Advances in machine learning and artificial intelligence have sparked interest from governments that would like to use these tools for predictive policing to deter crime. Early efforts at crime prediction have been controversial, however, because they do not account for systemic biases in police enforcement and its complex relationship with crime and society.

Data and social scientists from the University of Chicago have developed a new algorithm that forecasts crime by learning patterns in time and geographic locations from public data on violent and property crimes. The model can predict future crimes one week in advance with about 90% accuracy.

In a separate model, the research team also studied the police response to crime by analyzing the number of arrests following incidents and comparing those rates among neighborhoods with different socioeconomic status. They saw that crime in wealthier areas resulted in more arrests, while arrests in disadvantaged neighborhoods dropped. Crime in poor neighborhoods didn't lead to more arrests, however, suggesting bias in police response and enforcement.

[...] The new model isolates crime by looking at the time and spatial coordinates of discrete events and detecting patterns to predict future events. It divides the city into spatial tiles roughly 1,000 feet across and predicts crime within these areas instead of relying on traditional neighborhood or political boundaries, which are also subject to bias. The model performed just as well with data from seven other U.S. cities: Atlanta, Austin, Detroit, Los Angeles, Philadelphia, Portland, and San Francisco.

This strikes me as something more useful for determining where to improve crime prevention strategies (better lighting, etc.) than it is for catching perps in action.

Journal Reference:
Rotaru, V., Huang, Y., Li, T. et al. Event-level prediction of urban crime reveals a signature of enforcement bias in US cities. Nat Hum Behav (2022). DOI: 10.1038/s41562-022-01372-0

This AI Algorithm Supposedly Predicts Big City Crime Before it Happens. Is That a Good Idea?

Claims to predict future crimes a week in advance with 90% accuracy. So much for free will and all that ...

https://www.fastcompany.com/90766642/this-ai-algorithm-supposedly-predicts-big-city-crime-before-it-happens-is-that-a-good-idea


Original Submission

 
This discussion was created by janrinok (52) for logged-in users only, but now has been archived. No new comments can be posted.
Display Options Threshold/Breakthrough Mark All as Read Mark All as Unread
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
  • (Score: 3, Interesting) by Immerman on Sunday July 10 2022, @07:13PM (4 children)

    by Immerman (3985) on Sunday July 10 2022, @07:13PM (#1259558)

    I suspect your carjacking example is probably roughly correct - they might not even predict that it will be a carjacking, but just an "incident" of some sort.

    I don't know that even that is useless though - not unless there's an incident there most weeks.

    If you can generate a predictive "heat map" of where crimes are most likely to occur within the next week, you can make sure there's officers close at hand.

    The interesting bit might be to see what effect visibly patrolling those predicted hot spots might have. Would the crimes simply move elsewhere, or would that visible reminder of consequences actually reduce crime rates overall?

    Starting Score:    1  point
    Moderation   +1  
       Interesting=1, Total=1
    Extra 'Interesting' Modifier   0  
    Karma-Bonus Modifier   +1  

    Total Score:   3  
  • (Score: 2) by FatPhil on Monday July 11 2022, @09:47AM (3 children)

    by FatPhil (863) <pc-soylentNO@SPAMasdf.fi> on Monday July 11 2022, @09:47AM (#1259713) Homepage
    There already are heat maps of crime - I don't see what new thing this AI brings to the table.
    --
    Great minds discuss ideas; average minds discuss events; small minds discuss people; the smallest discuss themselves
    • (Score: 2) by Immerman on Monday July 11 2022, @09:10PM (2 children)

      by Immerman (3985) on Monday July 11 2022, @09:10PM (#1259914)

      There are heat maps of the crime that *has already* occurred. That's good for analyzing past performance, and at low enough space-time resolution, long-term trends.

      A heat map of where crime will occur tomorrow, or next week, lets you preemptively allocate resources more optimally, so that you can get more accomplished.

      • (Score: 2) by FatPhil on Tuesday July 12 2022, @07:24AM (1 child)

        by FatPhil (863) <pc-soylentNO@SPAMasdf.fi> on Tuesday July 12 2022, @07:24AM (#1260074) Homepage
        But that's what heat is - it's a statistical property. If you have no reason to conclude things will be different tomorrow, which would be an extraordinary claim that would require extraordinary evidence, the statistics of tomorrow are the statistics of today. So you do have a heatmap for tomorrow, it's the same heatmap you already have in your hands.
        --
        Great minds discuss ideas; average minds discuss events; small minds discuss people; the smallest discuss themselves
        • (Score: 2) by Immerman on Tuesday July 12 2022, @09:20PM

          by Immerman (3985) on Tuesday July 12 2022, @09:20PM (#1260296)

          If all their AI is predicting is that yesterday's statistics will repeat, then it's worthless and there's no story here.

          If it can actually make useful predictions, then it can generate a predictive heatmap of tomorrows expected activity that is different than today's actual activity.