Predictive Policing for Banksters

Schönes Ding von Sam Lavigne, Predictive Policing für Wirtschaftskriminalität im Finanzsektor, 'ne Robocop-App für Bankster.

White Collar Crime Risk Zones uses machine learning to predict where financial crimes will happen across the U.S. The system was trained on incidents of financial malfeasance from 1964 to the present day, collected from the Financial Industry Regulatory Authority (FINRA), a non-governmental organization that regulates financial firms.

The system uses industry-standard predictive policing methodologies, including Risk Terrain Modeling and geospatial feature predictors, which enables the tool to predict financial crime at the city-block-level with an accuracy of 90.12%.

Mic: This app exposes the white-collar criminals all around you

"The truth is that all of those systems use biased data sets and reinforce precisely the biases that they emerge from," Sam Lavigne, a technologist and the New Inquiry's special projects editor who helped develop WCCRZ, said in a phone interview on Tuesday. "And that the police departments making use of those systems work behind a veil of objectivity."

So the WCCRZ team, which included Lavigne, the New Inquiry publisher Francis Tseng and data scientist Brian Clifton, built out a "Most Likely Suspect" feature, scraping the LinkedIn profiles of top executives in that area to come up with a facial average. The app slaps a new face on crime: a thousand algorithmically generated, grinning white fraudsters.

It may not be the face of crime trotted out on the evening news, but that's the point. The app reframes the common cultural code for criminality — typically black, poor and uneducated — around corporations and neighborhoods that more accurately reflect where crime happens and who's committing it.