White Collar Crime Risk Zones

Through the New Inquiry-supported White Collar Crime Risk Zones (WCCRZ), contemporary artists Sam Lavigne and Francis Tseng, along with data scientist Brian Clifton, illustrate a paradox that lies at the intersection of data and law enforcement. It presents the satirical flipside of predictive policing programs like PredPol, which tend to heavily monitor street crime in low income neighborhoods, while turning a blind eye to “high level financial crime” committed by the likes of corporations and banks. Police departments around the United States rely on such programs to optimize operations, which analyze and learn from historical data to predict where crime is likely to occur and send cops there to prevent them. Some even go as far as to predict who is likely to be a victim. But the data fed into these predictive machines reflects how law enforcement have worked based on racist and classist assumptions in the past, and thus perpetuates the same uneven policing in an egregious feedback loop. Disguised as factual, efficient, and unbiased, predictive policing builds human prejudice right into itself.

Description of PredPol’s excellence, according to PredPol.

Thus, White Collar Crime Risk Zones proposes to pick up the slack here, where other predictive policing products drop it. There are three components to WCCRZ: A web application, an iOS application, and a technical white paper that explains the methodology of the White Collar Crime Early Warning System (WCCEWS), the machine that powers the applications. Taking on a satirically factual tone, the white paper presents WCCEWS as an opportunity to expand—not correct—predictive policing. Referring to academic research done on predictive policing and public surveys, WCCRZ identifies the gaps in the field and proposes to fill them with a model with admitted similarities to HunchLab, another predictive policing program. WCCRZ does not explicitly say why white collar crime—instead of, say, cybercrime—should be the next focus for predictive policing, but instead it appears to be presented simply as an untapped market. On one hand, this presentation takes away from WCCRZ’s obvious point that predictive policing is unfair and unjust, but on the other hand, it seems to show a convoluted support for predictive policing by saying why not be unfair and unjust to everyone then?

WWCRZ predicts how the possible perps may look based on averaged images of corporate executives pulled from LinkedIn.

The interfaces of the two versions of WCCRZ are quite similar. Both the web and mobile apps overlay Google Maps with predictive data that the WCCEWS algorithm churns out. The maps use color to indicate crime risk in different zones, from yellow to orange to bright red, corresponding with the severity of the crime. Places where risk of white collar crime is predicted to be particularly high aren’t surprising. The streets of Midtown, Manhattan, for example, are hardly visible under all of the bright red risk. In contrast, Governor’s Island has one measly yellow square, indicating an 80% chance of crime involving up to $10,000. Both applications also employ a tongue-in-cheek beta facial analysis feature, which is the averaged product of “the pictures of 7000 corporate executives whose LinkedIn profiles suggest they work for financial organizations.”


White collar crime risk zones around The New School.

While everything else about WCCRZ makes sense to me and satisfies my need for satire, the one thing that I feel slightly dilutes or confuses an otherwise fantastic project is its mobile component. For one, it is accessible for free through the AppStore. Furthermore, on mobile, unlike on web, the user can choose to get notifications whenever they are in a risk zone by setting the percentage of risk at which they want to be notified. By that standard, WCCRZ is not very different than my mobile banking app which lets me know when I’m near a branch. It is not obvious to me, though the white paper calls it a tool for “citizen policing and awareness,” how a mobile app available to the public supports the point against predictive policing—isn’t the whole critique that these technologies are mainly available to government agencies who can afford them? Perhaps in the name of truly committed satire, it means whatever you want it to mean, but to me it feels like a bit of an afterthought.

Relevant mobile app or not, WCCRZ visualizes a salient point not only about the way our cities are policed, but also how deeply our lives can be affected by the importance and trust placed on data. It reminds us that data and algorithms are not exempt from human prejudice—they are carriers of it.