Computing Fear in Black and Brown Communities
by Michael Dean Thompson
Over 50 years ago, fear of crime was even then associated in the minds of governing bodies with Black and brown communities. An effort to combat crime based on that fear spurred the creation of software that has since grown to become the predictive policing systems now known as fusion centers. Few, if any, valid metrics can justify such immense expenditure. Nevertheless, billions are spent annually across the nation to enhance predictive policing with invasive surveillance via cameras, facial recognition, automated license plate readers, cell-site simulators, and more.
It all began in 1966, not quite one year after President Johnson established the President’s Commission on Law Enforcement and the Administration of Justice, when he added the Science and Technology Task Force to the Commission in an effort to combat crime. At the head of the task force, they appointed Samuel L. Gass who at that point managed all of IBM’s federal system’s projects. As a result, the Police Beat Algorithm was created. The goal in creating the software was to determine how to divide a municipality based on demographics and geography in order to achieve an effective deployment of police resources. In direct response to recent events, including the Watts riots, Gass often pointed to the need to create “contingency riot and other emergency plans.”
The Police Beat Algorithm attempted to compute crime patterns by creating arbitrary weighting scales. Burglary, larceny, and auto theft received the same weight of four given to homicide and rape. Meanwhile, traffic accidents got a score of two and drunkenness managed a one. That they justified weighing a traffic accident that presumably had no culpable intent greater than drunkenness reflects the task force’s biases about the nature of criminality. Likewise, weights were given to geographical boundaries based on the amount of crime within them. Criminality was also normed within the boundaries based on particular groups such as white and Black residents. That is, they attempted to determine how much crime is common for each group in that area for each crime type, and therefore, what levels would represent unusual criminality.
Considering the technology available to them at the time, their goals were audacious. The task force wanted to create a system that could identify patterns within the crime data, associate those patterns with given suspects, attach suspects to past crimes, and predict where to place police resources. These were tremendous requirements asked of computers that pale in comparison to the modern cellphone and into which data was fed through punch cards.
Weighted geography and criminality meant that profiling was built into the system. Black and brown communities were preset to receive the lion’s share of police attention. For communities already suffering from aggressive over-policing came a scientific-seeming rationale to dedicate even more police rather than finding new solutions. The Police Beat Algorithm merely provided pseudo-scientific justification for existing biases. It was intended as a proof of concept, but in 1968, the Kansas City Police Department put it to use, and the impact of its implicit racism became evident.
The same biases that fed into the Police Beat Algorithm and lent it the thin veneer of scientific credibility infect modern predictive policing systems today.
As a digital subscriber to Criminal Legal News, you can access full text and downloads for this and other premium content.
Already a subscriber? Login