PredPol: A Case of Mistaken Identity
We are at a pivotal moment in this country where policing is being scrutinized for everything from racial discrimination, brutality, over-funding, and lack of skills training. Tools used by police departments are also under scrutiny. Santa Cruz City was the first city to ban predictive policing:
“Understanding how predictive policing and facial recognition can be disproportionately biased against people of color, we officially banned the use of these technologies in the city of Santa Cruz,” Mayor Justin Cummings.
The resolution by Santa Cruz City Council:
The purpose and intent of this Chapter is to prohibit the City’s acquisition and/or use of Face Recognition Technology and Predictive Policing Technology, prior to obtaining City Council approval, by resolution, based on the City Council’s finding that the technology meets scientifically validated and peer reviewed research, protects and safe guards the civil rights and liberties of all people, and will not perpetuate bias.
And to the surprise of many, Brian MacDonald, Santa Cruz-based predictive policing company PredPol, supports and commended our city leaders:
“Given the racial inequalities pervasive throughout American history and society, we as a company support this language. In fact, we would even go so far as to recommend that this standard be applied to all technologies adopted by the city of Santa Cruz, whether used for law enforcement purposes or not.” Brian MacDonald CEO PredPol
You might ask: HOW can a leader in predictive policing be in favor of such policies? Let’s look at the facts.
PredPol does not use facial recognition
Racial profiling has been associated with facial recognition. Consequently, PredPol does not use facial recognition. Period. Their algorithms use only 3 data points: event type, event location, and event date/time. Call it the what-where-when data. From the inception of PredPol, the engineers and management were clear about creating a tool that did not have racial or gender targeting. Under event type, they rejected using crimes associated by officer-initiated stops that might indicate bias or lead to profiling and discrimination such as: stop-and-frisk, drug sales / possession, gang activity, prostitution, gabling, sex crimes, crimes associated with minors, and domestic violence.
PredPol is the only predictive policing company that has no secrets
Still think that PredPol might have a dark secret under their algorithm? Fear not. Here it is:
Their patented algorithm public for all the world to see, test, and even improve upon. With other predictive policing companies, fat chance you will get to look under the hood. Who are those other companies to turn your facial recognition on? Palantir, IBM, LexisNexis, Cisco, Motorola, Shotspotter, Tyler Technologies, Hitachi Data Systems, and Splunk. LexisNexis even uses credit reports in their predictive policing. Help me Obi Wan.
The policy set by Santa Cruz City Council will have a negligible impact on PredPol. Only about 10% of the PredPol product is focused on predictive capabilities; the rest is used for officer location tracking and crime analytics. When and where it has been used, it has had very positive results. Over a 5-year period from 2013 to 2018, the city of Modesto, California, reduced vehicle thefts by 45% and residential burglaries by 75%. The majority of PredPol customers use the software for tracking police activities. Management can track and archive daily police activities, which reveal insights such as when police are spending too much time in one neighborhood needlessly. And isn’t this what we want: transparency?
For an excellent article digging deeper into the misunderstanding about PrePol, read Jacob Piece “Was It Racist? Santa Cruz Bans Predictive Policing” in the Good Times.