The Benefits and Implications of AI in Law Enforcement

The Belle Fourche Police Department came up with a plan to integrate cameras into its operations, and it implemented the plan in four phases. The first phase was setting up the AI-enhanced cameras throughout the city. The second phase involved positioning the cameras in high-traffic areas, such as downtown and on bike paths.

“Once we had that in place, we saw our property and vandalism calls drop from just over 1,000 to just over 200 in the first six months,” Cherveny says. “And it’s stayed there ever since.”

For phase three, the police department placed cameras in all city buildings and city properties. It is now in phase four, which entails placing cameras in high-crime areas and at city intersections.

“Utilizing that technology has drastically cut back on our calls for service and sped up our response times,” Cherveny says. 

AI Solutions Boost Police Efficiency

Verkada Public Safety Advisor Ben Jones formerly worked as a lieutenant with the Winston-Salem Police Department in North Carolina. He recounts using Verkada AI-enhanced cameras in the city’s real-time crime center.

“It helps to aggregate a lot of data really quickly,” Jones says.

For example, the Winston-Salem Police Department could identify vehicles suspected of involvement in homicides or robberies within 15 to 20 minutes thanks to a combination of the camera platform and AI flagging the footage of interest. Previously, a human would have to spend hours watching the videos in order to find a vehicle.

The police department also used AI to quickly recall relevant records. If officers received a call about a domestic dispute, for example, they could compare the address or the names of people involved in the call against previous police encounters in order to help determine the intensity of the threat officers might face in an encounter. Officers can get that vital information quickly thanks to AI, Jones says.

“The evolution of people in that room coordinating and learning together and knowing what’s good information and what’s not good information and getting that information out has been the real power of the real-time crime center and the analytics that go with it,” Jones says.

AI in the cloud also supports a common operating picture between operators in different locations, which is critical to successful police work, Jones says. He recalls an incident in Winston-Salem where the real-time crime center and a school couldn’t share information, meaning it took longer to understand the situation on the ground. In a different incident at a community college, the real-time crime center and the college synchronized video immediately, and they were able to share a common view quickly.

“I can work from the real-time crime center, the command post can work with me, and we are all seeing the same common operating picture,” Jones says.

EXPLORE: Smarter real-time crime centers help solve cases and secure communities.

Law Enforcement Earns Public Trust for AI Applications

Dean Cunningham, segment development manager, public safety for Axis Communications, started the body-worn camera program for the police department in Fort Collins, Colorado, in 2012, when he was an officer there.

Body-worn cameras give a new level of visibility into police work, which increased public confidence in the police force, Cunningham says. With the introduction of AI transcription services, body-worn cameras can now transcribe everything that happens in officer interactions, so everyone can read the exchanges.

“It changed how people look at how the jobs get done,” Cunningham says. “The cameras give you an opportunity to learn.”

To increase public trust in the use of any new technology, law enforcement agencies should communicate directly with the public about solutions and how they are being used. “Your public information officer should talk to people about what the technology is and what the technology is not," Cunningham says.

The public will also have more confidence in AI solutions if it’s obvious that a human being oversees the AI, he adds.

Click the banner below to sign up for the StateTech newsletter for weekly updates.

Jones agrees, noting that the Winston-Salem Police Department opened the doors of its real-time crime center to the public upon request so people could see how police were working. “The worst thing you can do is put out bad information. That is going to ruin your credibility,” he says.

The Winston-Salem Police Department real-time crime center also involved the community by integrating volunteered cameras into its operations. More cameras expanded the power of the real-time crime center while also gaining community buy-in. The police department could then audit who was doing what, and how, with Verkada software granting even greater transparency.

Belle Fourche Chief of Police Ryan Cherveny says the public gains confidence in AI technology if you explain clearly how it is used while respecting the privacy of citizens. His town’s cloud-based system only retains video and photos for a short period of time before they are deleted.

“It is not saved in a database; it is there for a certain period of time and then it is gone,” Cherveny says.

The Future of AI in Law Enforcement

In Belle Fourche, police were able to find a missing child because they could map his travels across town through cameras that captured him at different locations. AI verified the child was the same person in the various locations throughout a day. 

Cherveny would like to see AI systems get smarter when it comes to mapping the location of people and vehicles in similar ways.

“AI produced a map and showed us where they were from one spot to the next,” the chief says. “Enhancing and broadening that capability is something that I look forward to."

Jones says analytics will continue to improve in the future, helping the chief and others gain the kind of visibility they seek. Verkada, for instance, would like to get more detail on vehicles and empower AI to track those details. Does a vehicle have a dent on it? Does it have a sticker? Can we track a vehicle based on these details with a high level of reliability?

“We live in an age of intelligence-led policing,” Jones says. “We want to provide as much intelligence upfront before an officer arrives. A more informed officer is going to make better decisions.”

Officers want open systems, Cunningham observes. Different police departments work in different ways, and they want interoperability for camera systems and software across their environments.

To this end, AI solutions providers must be sensitive to how their products can scale while also working well with other solutions in jurisdictions’ tech stacks.

“The needs of a police department in Tulsa, Okla., are different from those in Houston, Texas,” Cunningham says. “Everyone needs to use what works in their processes.”