Artificial intelligence (AI) has become a major player in policing, and it is likely here to stay. And, as police retirement numbers increase while recruitment numbers plateau, embracing AI can help cover gaps in services, allowing agencies to more effectively allocate resources.

But as the technology’s capabilities continue to expand in the field, the loss of human oversight becomes a point of worry , creating an “accountability gap.”¹ Foreseeing this concern, several companies have created their products with the human at the heart of every decision. “Humans and AI are exponentially more powerful together,” said Hamish Dobson, corporate vice president of enterprise physical security at Motorola Solutions. “AI should always complement—not replace—human decision-making.”²

Data Analysis Tools

From body-worn camera footage to information gathered from social media sites, there are more data to analyze than ever. “Because of the vast amount and diversity of data available today, it has become overwhelming for analysts and investigators to leverage everything available to them,” said NEXYTE’s Vice President of Sales Mark Massop.³ Collecting and analyzing these data without the help of AI takes precious agency resources, human resources, and investigation time.

Cognyte’s decision intelligence platform, NEXYTE, scopes available data sources and information to determine important data from an investigative standpoint. “By embracing technology like NEXYTE, agencies can help to cover the gaps in services, using data analytics to better understand where and when to deploy those limited resources,” said Massop. The developers at Cognyte worked with advanced machine learning models to accelerate the investigative process by allowing the computer to analyze mass amounts of data and surface new leads to empower the investigator. The solution provides a suite of built-in machine learning algorithms to discover, prioritize, and link potential suspects.

Deployed on the client’s hardware or cloud-hosted environment, relevant data are imported to the NEXYTE platform where users can interact with the information in a single-pane-of-glass view. “Only by bringing the data into a single repository can you harness the power of AI and [machine learning] to accelerate the decision-making process,” said Massop.

The decision intelligence platform automatically links relevant data points related to individuals, groups, or objects as part of the entity resolution process. This action is performed while the enrichment layer uses AI to extract and correlate metadata from structured and unstructured sources. Once the data are arranged, NEXYTE can apply additional analytics to gain automated insights.

“Because it is difficult for law enforcement and other security personnel to monitor hundreds of video security cameras in different locations, Motorola Solutions created analytics that direct a technician’s attention to where they need to focus,” said Dobson.⁴ This technology can be found within Motorola Solutions Avigilon Unity security suite’s Unusual Activity Detection (UAD), which allows law enforcement to investigate and direct resources without having to manually monitor every security camera. Alerts generated from UAD technology appear through the Avigilon Unity Focus of Attention user interface, which highlights areas that require attention.

For example, imagine if a vehicle sped in a slow-speed pedestrian zone or if someone climbed onto a low height structure within range of a security camera. A technician monitoring the video security cameras in a real-time crime center would be alerted to the unusual activity, and the analytics program would direct the technician’s attention to the appropriate camera. Officers can quickly review these alerts to determine whether a response is needed.

Hexagon’s Safety, Infrastructure & Geospatial division solutions also work within the real-time crime center environment. Embedded within the HxGN OnCall Dispatch solution and the HxGN Connect solution, its assistive AI (also known as Smart Advisor) acts as a digital detective within a public safety agency’s computer-aided dispatch system and real-time crime center to provide actionable insights during complex emergencies.

Crime analysts can set rules to notify dispatchers and investigators if a specific incident has occurred. Smart Advisor then combs through data and live incident information to identify keywords, trends, and anomalies. Once an event is created, a dispatcher will receive recommendations on the next steps to take based on the rules that the analyst input. The dispatcher can alert detectives and attach any of Smart Advisor’s recommendations to the incident record without leaving the dispatch environment. The detective will receive this message on his or her computer-aided dispatch device and can view the event details without needing a Smart Advisor license. After reviewing the alert, the detective can decide if and what action is necessary.⁵ “Smart Advisor, as the name suggests, is purely an advisor to the user, offering additional insights to support informed decision-making,” said Bill Campbell, senior vice president of global public safety at Hexagon. “The user is always in control.”⁶

Conclusion

At the heart of every machine are the humans—the developers and the users. Although there are concerns that the lack of human oversight will cause bias and inaccuracies, companies have developed these products with human decision-making in mind. No decision is made by AI, the user must confirm all decisions.

As veteran supervisors, call takers, and dispatchers are looking toward retirement, agencies could lose valuable institutional knowledge that came from years of experience in manually connecting the dots in real time. AI technology and software can allow agencies to retain that institutional knowledge over time. Combining human experience with assistive AI can provide agencies with the crucial insights needed to become more efficient, aware, and proactive. d

Notes:

¹Hope Reese, “What Happens When Police Use AI to Predict and Prevent Crime?,” JSTOR Daily, February 23, 2022.

²Hamish Dobson (corporate vice president, enterprise physical security, Motorola Solutions) email interview, May 9, 2023.

³Mark Massop (vice president of sales, NEXYTE), email interview, April 21, 2023.

⁴Dobson, email interview.

⁵Hexagon, “Detect Complex Emergencies Sooner With the Help of Assistive AI.”

⁶Bill Campbell (senior vice president, Global Public Safety, Hexagon’s Safety, Infrastructure & Geospatial division) email interview, May 2, 2023.