comScore
  1. Mediaite
  2. Gossip Cop
  3. Geekosystem
  4. Styleite
  5. SportsGrid
  6. The Mary Sue
  7. The Maude
  8. The Braiser

What's with the name?

Allow us to explain.

The Future Is Now!

Minority Report Precog-Like Software Being Tested In Baltimore And Philadelphia To Predict Crimes


More and more these days we’re witnesses the technologies from our favorite sci-fi films and television shows get closer to reality. The latest is a new kind of software meant to predict the actions of criminals. Baltimore, Maryland and Philadelphia, Pennsylvania are already using it to prevent murders and Washington D.C. is next on the list. Don’t worry, no one is keeping psychics in a pool. Well, not that we’re aware… 

The technology was developed by Richard Berk, a professor at the University of Pennsylvania. The hope is a decrease in murder and other crimes but how does it work? The software isn’t used to predict what you or I might do in the future but what those already convicted of crimes and headed for parole might do.

“When a person goes on probation or parole they are supervised by an officer. The question that officer has to answer is ‘what level of supervision do you provide?’” Berk told ABC News. The level of supervision is dependent on the parole officer’s best guess essentially. They look at the prisoner’s record and decide what they feel is appropriate but now, it’s the software that assesses the likelihood to commit another crime.

Beginning several years ago, the researchers assembled a dataset of more than 60,000 various crimes, including homicides. Using an algorithm they developed, they found a subset of people much more likely to commit homicide when paroled or probated. Instead of finding one murderer in 100, the UPenn researchers could identify eight future murderers out of 100.

Berk’s software examines roughly two dozen variables, from criminal record to geographic location. The type of crime, and more importantly, the age at which that crime was committed, were two of the most predictive variables.

So it’s not simply a case of a murderer likely to commit the crime again but many factors.

“People assume that if someone murdered then they will murder in the future,” said Berk. “But what really matters is what that person did as a young individual. If they committed armed robbery at age 14 that’s a good predictor. If they committed the same crime at age 30, that doesn’t predict very much.”

The software is also used to predict which criminals may themselves be the victim of murder and leads many to make the connection to the Steven Spielberg film, Minority Report. Berk said, ”We aren’t anywhere near being able to do that.” Washington, D.C. is currently only testing the software on lesser crimes but if proven successful will likely spread out its use to more serious offenses.

(via Wired)

Are you following The Mary Sue on Twitter, Facebook, Tumblr, Pinterest, & Google +?

TAGS: | |


  • http://www.facebook.com/ronald.stepp Ronald Stepp

    “The level of supervision is dependent on the parole officer’s best guess essentially.”

    “So it’s not simply a case of a murderer likely to commit the crime again but many factors.”

    Wow, sounds really sciencey.

  • Anonymous

    Honestly, this sounds completely horrifying and exploitable, an abuse of science by the government. Don’t want to let someone out of prison? Oh, well, The Algorithm says you might be likely to commit a crime, so let’s just keep you locked up on the freaking off-chance. Sheesh. As if the system isn’t broken enough already.

  • ACF

    I’m not sure I see how you think this worsens the system any; if the system is corrupt, officials can keep a prisoner in jail longer just as easily without the software (probably more easily, in fact; bear in mind that the previous system involved the parole officer’s best guess, which would be much easier to change than the result of a software algorithm).

    None of this is to say that there aren’t current problems with our prison system, but at least the software appears to make the current system function a little better, by offering better predictions.

  • http://twitter.com/JMCanuck Jim

    The area of study is called “Predictive Analytics” and if you’ve every applied for life insurance, an underwriter would have used a more primative technique of evaluating your fitness for it and determined the rate you pay. Age, gender, smoker (or not) are the basics, more factors if it required medical review. It happens.

    Now, I’m just gonna say, used for good, a parole officer could help steer a parolee toward choices that would reduce his risk. (Much the same as your doctor might say losing that spare-tire’s gonna make you live a decade longer.)

    Data is never evil, it’s how it’s used.

  • Jesse

    Sounds like more funding to punish criminals instead of stopping crime and helping people. I agree this could be used to steer people who are ‘at risk’ in the right direction – but i don’t see that happening when convicts and ex-convicts are second class citizens with very few rights and are routinely legally discriminated against. If the government would just fund schools, improve the quality of education, fund projects to lower poverty, then crime would lessen. Instead we focus on punishing people because we don’t actually want to lessen crime because prisoners provide free slave labor, because poor people are too busy dealing with stress and working all the time to organize and democratize our society.

    Yes I am ranting but seriously, few things anger me like the prison system in the US.

  • Anonymous

    I agree the current system is broken. But as difficult as it is to surmount the current joke of a parole system, it’s a lot easier to argue against one parole officer’s admitted opinion than against what is perceived to be infallible science. I don’t mean to belittle science here, only people: When people see a scientific result and a mathematical formula that feels right and gives them an easy answer and eases their consciences, they turn off their brains and go with it.

    8 out of 100 vs. 1 out of a 100 – the 8 might be better, but it’s still not very good, and there are still way too many false positives to make me happy.

  • ACF

    I think you might be misunderstanding the results, or maybe I’m misunderstanding you; it says they can predict 8 murderer out of 100, meaning of 100 murderers, it predicted 8 of them, not it predicted 100 people would commit a murder, and 8 actually did. The rest of your objection may or may not be valid, depending on how it’s applied; I don’t know enough about how current recommendations are reviewed to say.

  • Anonymous

    Soooo….Batcomputer?