40% off TNW Conference!
Residents of neighborhoods where PredPol suggested few patrols tended to be Whiter and more middle- to upper-income.
Many of these areas went years without a single crime prediction.
These communities werent just targeted morein some cases they were targeted relentlessly.
A few neighborhoods in our data were the subject of more than 11,000 predictions.
They need resources to fill basic social needs.
The same disparity existed between richer and poorer communities.
This isnt a continuation of research.
Its like trying to diagnose a patient without anyone fully telling you the symptoms, Musa said.
The prosecution doesnt say, The tool that we purchased from this company said we should patrol here.
None of it matched perfectly with arrest reports during that period, which were also provided by the agency.
Public access to that page is now blocked.
That left 5.9 million predictions provided to 38 agencies over nearly three years.
Company CEO Brian MacDonald said our data was incomplete, without further explanation, and erroneous.
After we declined, he did not respond to further emails.
One exception was the Decatur Police Department in Georgia.
John Bender said in an emailed statement.
We asked MacDonald whether he was concerned about the race and income disparities.
MacDonald said the company didnt adjust its software in response.
Such a change would reduce the protection provided to vulnerable neighborhoods with the highest victimization rates, he said.
PredPoladvisesofficers to get in the box during free time.
Officials in some cities said officers frequently drove to prediction locations and completed paperwork there.
Who Reports Crime?
We use crime data as reported to the police by the victims themselves, he said.
If your house is burglarized or your car stolen, you are likely to file a police report.
But thats not always true, according to the federal Bureau of Justice Statistics (BJS).
In aspecial reportlooking at five years of data, BJS found an income pattern as well.
This disparity in crime reporting would naturally be reflected in predictions.
There is only reported crime data.
And the difference between the two is huge.
Five cities provided us with data on officer use of force, and we found a similar pattern.
In Piscataway, N.J., the arrest rate was more than 10 times the city average in those neighborhoods.
And now they have the data to prove it.
Take the 111-unit Buena Vista low-income housing complex in Elgin.
Six times as many Black people live in theneighborhoodwhere Buena Vista is located as the city average.
Brianna Hernandez had spent two years on a waiting list to get into Buena Vista.
It was November 2020.
You know youre not supposed to be here, right?
King remembers Miller asking him.
They told him that once you got off probation you would be able to come back, Hernandez said.
Apparently, that didnt happen.
It was Kings third arrest for trespassing at Buena Vista.
Miller arrested him for trespassing and weapons possession.
That case is still pending.
I know hes banned, but what can a man do?
She said the arrest led to the eviction notice from Buena Vista.
(Buena Vista wouldnt confirm or deny it.)
Hernandez remembers her 4-year-old and 5-year-old children asking, Why are we going to a hotel?
and struggling for an answer.
They want to know why were moving stuff out.
Why this and why that….
I wanted to sit down and cry.
We felt like these kinds of design decisions mattered, he said.
Advocates we spoke to in at least six cities were unaware PredPols software was being used locally.
Even those involved in government-organized social justice committees said they didnt have a clue about it.
We found four municipalities used PredPol to predict drug crimes between 2018 and 2021: Boone County, Ind.
The other two were Birmingham and Fort Myers, Fla.
If they decide to add other event types later that is up to them.
The other agencies declined to comment about it or ignored our questions altogether.
Nearly every agency also combined fundamentally different crime types into a single prediction.
MacDonald said research and data support the fact that multiple crime types can be concentrated in specific crime hotspots.
Christopher Herrmann, a criminologist at the John Jay College of Criminal Justice, disagreed.
Crime is very specific, Herrmann said.
The serial shoplifter isnt going to start stealing cars.
A serial rapist isnt going to start robbing people.
The officer ran Mosess name and discovered he had a warrant for an unpaid fine for fare evasion.
Moses was cuffed, searched, and thrown in jail for the night.
you’re able to just be at the wrong place at the wrong time.
The LAPD didnt respond to questions about whether the officer was responding to a PredPol prediction.
We did not have a go at determine how accurately PredPol predicted crime patterns.
Its main promise is that officers responding to predictionsprevent crimes by their presence.
These include Piscataway; West Springfield, Mass.
Craig Koostra said in a written statement.
He did not respond to a request to elaborate.
Some agencies soured on the software quickly.
Even PredPols original partner, the LAPD, stopped using the software last year.
The departmentsaidit was a financial decision due to budget constraints.
And he said he learned a lot at their feet.
Criminals are effectively foragers, he added.
Choosing what car to steal is like choosing which animal to hunt.
Among the signatories were 13 professors, researchers, and graduate students at UCLA.
MacDonald in turn criticized the critics.
And that, too, was Fergusons point.
Theyre just not going to call it predictive policing, he added.
And its going to be harder to pull apart for journalists and academics.
