Policing in the era of big data

Kathleen McGrory and Neil Bedi

July 21, 2021

Key findings:

  • The Pasco County Sheriff's Office generates lists of people it considers likely to break the law based on their arrest histories, unspecified intelligence and arbitrary decisions by police analysts. 
  • The agency sends deputies to monitor and harass those individuals, often without probable cause, search warrants or evidence of a crime.
  • In five years, the program ensnared almost 1,000 people. At least one in 10 were minors.

Key findings:

  • Separately, the Sheriff's Office creates a list of schoolchildren it considers likely to become future criminals based on factors including their grades, school attendance records and child welfare histories. 
  • The Sheriff's Office doesn't tell the kids' parents or teachers about the designation. The schools superintendent initially said he was unaware district data was being used this way.
  • Last October, 420 middle- and high-school students were on the list.

Both initiatives fall under the agency's Intelligence-Led Policing program 

We later explained how the powerful and politically connected sheriff had created the intelligence machine. 

Impact:

  • Four people targeted by the Sheriff's Office filed a federal lawsuit against the agency with the support of a national public interest law firm.
  • Florida lawmakers proposed two bills to curb the policing tactics described in our series.
  • Civil liberties groups denounced the intelligence programs and promised to take action. Thirty national and state groups including the NAACP Legal Defense Fund and  Southern Poverty Law Center formed a coalition.

Impact:

  • The Sheriff's Office and the school district later announced that school resource officers would no longer have access to student grades, and police analysts would have to make a record whenever they accessed confidential student data for law enforcement purposes. 
  • At the urging of a U.S. congressman, the federal Department of Education opened an investigation into the data sharing between the Sheriff's Office and the local school district.

Squitieri et al v. Nocco (2019)

Our reporting:

  • While we made our case for the records (and then waited for them), we knocked on many, many doors. We logged information from each interview in a series of spreadsheets.
  • We asked for interviews with the sheriff and agency leaders in October 2019.
  • We requested dozens of public records, including target lists, dispatch logs, intelligence reports, grant applications, body-camera footage, employee rosters and budgets.

Our reporting:

  • We knew our model was missing some pieces of data. But it helped us identify patterns of policing and zoom in on certain case studies.
  • We used an internal manual to reverse engineer the formula for giving scores to residents. We applied it to a massive database we built that joined arrest records, dispatch logs, code enforcement citations and use-of-force reports.

Our reporting:

  • We spent time with eight individuals/families who had been targeted. To fully understand their stories, we reviewed hours of body-camera footage, police reports and in some cases, juvenile records.
  • We built an interactive graphic using body-camera footage that showed readers what it felt like to be targeted.
  • We identified and interviewed insiders who could help us understand the program's evolution and how it was being applied.

The agency's response:

"In statements that spanned more than 30 pages, the agency said it stands behind its program... It said other local departments use similar techniques and accused the Times of cherry-picking examples and painting 'basic law enforcement functions' as harassment.

 

"The Sheriff's Office said its program was designed to reduce bias in policing by using objective data. And it provided statistics showing a decline in burglaries, larcenies and auto thefts since the program began in 2011."

We published the full statements online, as well excerpts organized by topic. We fact-checked as necessary.

What's happening in your area?

  • Ask your local law-enforcement agencies how they are using data and algorithms to drive decisions.
  • Be transparent and fair.                                 
  • Request internal manuals that explain "intelligence" programs.
  • Request contracts with software companies like Palantir and PredPol.
  • Investigate other uses of AI in policing, including autonomous drones and robots, and facial recognition software. 

Policing in the era of big data

By Kathleen McGrory

Policing in the era of big data

  • 59
Loading comments...

More from Kathleen McGrory