Predictive policing brings burglary numbers down, but is privacy at risk?

Filed Under: Featured, Law & order, Privacy

image of a police officerLos Angeles police are the latest to embrace "predictive policing." This technique replaces colored pins on maps with databases that identify crime hot spots based on past crimes and patterns.

The new technology, first rolled out in the San Fernando Valley late last year, is being credited with double-digit drops in burglaries and other property crimes.

The Associated Press reports that over the last 10 years, large police departments, like New York City's and Los Angeles's, have been using the similar, less sophisticated CompStat system to track trouble spots.

The LAPD and police in Santa Cruz, Calif., are now using a new program claimed to be more timely and precise.

Here's how the AP story describes it:

Built on the same model for predicting aftershocks following an earthquake, the software promises to show officers what might be coming based on simple, constantly calibrated data — location, time and type of crime.

Writing for Police Chief Magazine, LAPD Chief of Detectives Charlie Beck and Colleen McCue, president and CEO of MC2 Solutions, compared the technology to Wal-Mart's just-in-time stocking of Pop-Tart pastries for stores located in the path of storms.

(Apparently, consumers in danger of storm shortages eat a lot of Pop-Tarts. Specifically, strawberry. Who knew?)

Rather than boost the number of arrests, the AP reports, the goal of predictive policing is to intercept a crime in progress or to deter would-be criminals.

The AP quotes Jeff Brantingham, an anthropology professor at the University of California, Los Angeles, who said the data is in part derived from the criminal behaviors of repeat victimization and the notion that criminals tend not to stray too far from areas they know best:

"If you are victimized today the risk that you'll be a victim again goes way up," said Brantingham, who co-founded a software company tapped by LAPD for its program.

It must be working: crimes declined in the predictive policing target area 13 percent following the software rollout, compared with a slight increase in the areas of the city where the program wasn't used.

What could possibly go wrong?

image of a thiefTo answer, we could take a look at Los Angeles's Suspicious Activity Reporting (SAR).

As Uzma Kolsy wrote for Salon, SAR has its roots in the 1970s and 80s, starting as "a large-scale domestic spy operation" that enabled Los Angeles police to snoop on and infiltrate political, labor and civic organisations, including the office of then Mayor Tom Bradley.

The LAPD might not go that far nowadays, but they're still empowered by a directive titled Special Order 1 that permits officers to dub something "suspicious" and then act on it.

Kolsy writes:

"Where things get murky … is how SAR guidelines categorize constitutionally protected, non-criminal and commonplace activities such as using binoculars, snapping photographs and taking notes as indicators of terrorism-related activity. The SARs are coupled with the LAPD’s iWatch program, a campaign the police pioneered to encourage regular citizens to report 'suspicious' activity, including 'a person wearing clothes that are too big or too hot for the weather,' or things that just plain old don’t 'look right.'"

Clothes that are "too big or too hot for the weather"? You can't read that without thinking of hoodies, which is reminiscent of incidents such as the tragic, fatal shooting of Trayvon Martin by George Zimmerman.

The National Institute of Justice holds symposia on predictive policing. During one, a panel on privacy and legal issues acknowledged that SAR holds lessons about the importance of establishing a privacy policy to help keep predictive policing from becoming a black hole for civil liberties.

Other conclusions from a break-out session on privacy:

  • There is a rich history dealing with privacy and mistakes, and these issues have yet to be resolved
  • There will come a time when training in privacy issues is considered as important to a policing program as the firearms policy
  • Transparency is critical to establishing community trust
  • Understanding the behaviors that have a nexus to crime provides a valid law enforcement purpose
  • Predictive policing must be constitutional.

OK. Fine. Good.

One thing: why aren't privacy issues as important as a firearms policy now?

sign: Private: keep outWhat's stopping police from making privacy issues a priority? The mere fact that privacy isn't given the same importance as guns should perhaps give us all pause when it comes to predictive policing.

If police want communities to embrace this technology-enabled initiative, they should make a good faith effort to train officers on privacy policy and beef up (or institute) anti-profiling training.

Image of police officer courtesy of Shutterstock
Image of neighbourhood courtesy of Shutterstock
Image of thief courtesy of Shutterstock
Image of Keep out sign courtesy of Shutterstock.

, , ,

You might like

7 Responses to Predictive policing brings burglary numbers down, but is privacy at risk?

  1. Andrew · 1192 days ago

    The article seems to be conflating two separate issues. Predictive modelling as described should produce useful results even in the absence of SAR - although the more information it's given, the better it should be able to segment the area based on statistically significant differences. A computer algorithm would be effectively blind to any ethnic or other prohibited factors, unless they are fed into it and found to be predictive. Unlike humans, who can be psychologically swayed by such things to a greater degree. So, the privacy issues seem to be all at the front end, regarding what data is collected and how it is described.

    The title of this article creates a false connection between the use of computer models for predictive policing and the collection of observational data. The privacy issues discussed seem to all relate to the latter, regardless of whether it's used for predictive modelling or simply traditional direction of police resources.

    • Katherine Anthony · 1192 days ago

      Couldn't have said it better. I don't have an issue with predictive modeling. Or rather, its nice to see it used for something other than trying to sell me junk and futures on diabetes. Maybe something that could even be good for society. The problem isn't that the technology is used to stop crime, it's that it could be used to stop things that *aren't* crime.

      Where I live, downtown is struggling to exist, partly because the police harass and often taser people trying to enjoy the public squares. They've got it in their heads that only drug users use the area, so anyone there must be a drug user. That mentality is the issue, and they've achieved that without any technological aid.

      • Paul · 1115 days ago

        I feel for you in you wish to be free of prejudice. The same issue arises when a black man drives through an affluent white neighborhood. There are several thoughts I have on this subject. 1) For those regular, real, and law abiding (you did not use this term) who want to take back the parks for kids and the community, I applaud you. but it's only by working with 'the man' and gaining the respect of 'the man' that he is likely to lower his expectations of what he has to do. It is too much in this day and age to assume that any of us is immediately entitled to respect. You have to work at it; and it doesn't come easy. 2) Regarding Predictive Modeling and SAR sorry but unless the data is faulty, if it shows the park as a prime drug area you still need to work with the authorities (probably quietly) to remove the bad influence from your children. 3) If you do not feel threatened, contact your politicians and ask them to see if your area could be tweaked back for the reasons you cite.

        There is a middle ground to come to for the protection and safety and security of all without locking down the park but it involves work, visibility and open-mindedness on both sides

  2. Guest · 1192 days ago

    "why aren't privacy issues as important as a firearms policy now?"

    ... because they are in direct opposition to predictive policing... that's why.

  3. Randy · 1192 days ago

    "databases that identify crime hot spots based on past crimes and patterns"

    These databases do not include information on individuals, just info on criminal events. This is not a privacy concern at all. If statistics show that a mugging occurs at 8th Ave. and Jackson St. between midnight and 2:00am three times a month, how does that information affect my privacy? Does this information identify the mugger? Does it give away his Social Security number? Does it reveal his/her medical history? The muggers might complain about increased police patrolling in the area but I certainly won't.

  4. mittfh · 1192 days ago

    I suppose a lot depends on what you do with the data generated by the predictive policing computer system. If it's just used to spread resources more efficiently by increasing patrols in known hotspots while reducing them in low-crime areas, I don't see an issue with it.

    However, if the police going on patrol are specifically told to be "extra vigilant" because although the area looks innocuous, it's a high crime area; that's when it could lead to over-zealous policing - "Oi! You look dodgy! You're nicked!"

    I deliberately said "area looks innocuous", as I'd imagine in most towns and cities, there's a public perception of crime hotspots (which may or may not correlate with those in the database), so even in the absence of the database may be extra vigilant when on patrol there anyway.

  5. Mary · 1191 days ago

    Am I reading correctly that "snapping photographs" is regarded as an indicator of potential illegal activity? In Los Angeles?

    How many tourists – or paparazzi – will be stopped and searched before it is realized that primary result of this scheme will likely be a massive waste of police time checking everyone carrying a camera, which must quickly become unsustainable regardless of short term reductions in crime statistics.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

About the author

I've been writing about technology, careers, science and health since 1995. I rose to the lofty heights of Executive Editor for eWEEK, popped out with the 2008 crash, joined the freelancer economy, and am still writing for my beloved peeps at places like Sophos's Naked Security, CIO Mag, ComputerWorld, PC Mag, IT Expert Voice, Software Quality Connection, Time, and the US and British editions of HP's Input/Output. I respond to cash and spicy sites, so don't be shy.