Police push back against using crime-prediction technology to deploy officers

The Burbank Police Department has suspended officer deployments based on “predictive policing” technology hailed by top brass as the future of crime-fighting after complaints from police officers.

The shift comes as police departments across the country are increasingly using computer technology to help predict crime trends and deploy officers accordingly. Some law enforcement agencies have said the system has helped crack down on crime.

But in Burbank, critics said the software’s algorithm couldn’t beat a veteran officer’s intuition and knowledge of his or her patrol area. They also said the algorithm sometimes zeroed in on obvious areas where officers already know there’s crime or silly locations, such as the police station, where people often show up to report crimes.

Roughly half of the 118 sworn employees who took a June survey by the Burbank police union indicated that they’d considered leaving the department sometime during the last year, while 60% indicated they were unlikely or extremely unlikely to recommend a career with the agency to their children, relatives or friends.

About 75% indicated that morale at the agency was low or extremely low.

Dozens advocated that the agency hire more officers, increase specialty assignments and rely less on the predictive policing technology, which analyzes crime reports to predict potential problem areas.

The survey results prompted several discussions with the department’s administration that union leaders hoped would improve working conditions.

The agency started using the predictive policing technology in late 2014 at an annual cost of $15,000.

The technology analyzes crime data — the location, time and type of crime — to predict crime patterns. Daily, it spits out 500-feet-by-500-feet “boxes,” or potential problem areas, in which officers were required to spend about 15 minutes during their 12-hour-20-minute shifts. The idea was that visibility in those areas would deter crime.

During a shift, officers were given up to three boxes, which would translate into roughly 45 minutes of time. Sometimes, they’d get none. They’re no longer required to spend time in the “boxes.”

“We’re still pushing the information out and telling people to pay attention to what’s being said,” Burbank Police Chief Scott LaChasse said.

He doesn’t have a problem with officers self-deploying based on their knowledge of their patrol areas but said, “The future in law enforcement is not going to be random patrol, it’s going to be predictive analytics.”

The crime analyst who oversaw the program recently left the city for another job. LaChasse said he will reevaluate how the department uses the technology once that vacancy is filled.

It’s not clear how the technology has affected crime in Burbank, which rose 9% last year.

The Los Angeles Police Department has been a leader in the use of predictive policing technology. The department uses heat maps, technology and years of statistics to identify crime “hot spots.” Helicopter pilots then use their downtime to fly over them, on the theory that would-be criminals tend to rethink their nefarious plans when there’s a police aircraft hovering overhead. Police have said crime fell during periods when the helicopter program was operating.

British researchers are working with the LAPD and Santa Monica-based Rand Corp. to monitoring millions of tweets related to the L.A. area in an effort to identify patterns and markers that prejudice-motivated violence is about to occur in real time. The researchers then will compare the data against records of reported violent acts to see if predictions can be made.

alene.tchekmedyian@latimes.com