AI Partners – Filling Law Enforcement Experience Gaps

The Marshall Project, a nonprofit, nonpartisan news organization about criminal justice, states that Americans are choosing to opt out of policing and other government jobs. According to their analysis, many officers resign or retire early after incidents such as the Floyd Protests in 2020-2021. Senior police officers are leaving the law enforcement profession. Imagine retaining most of the department’s experiences and wisdom, even deploying that institutional knowledge with every officer in the field. This ability could soon be as simple as using a device like the one used to read this article. Artificial intelligence (AI), though, could offer much more expertise to any contact in the field, which supports law enforcement efforts to implement AI and machine learning (ML) into the daily lives of officers on duty for better and safer communities. It is imperative to look at how improvements in policing involving AI would benefit citizens in the short term and move forward. 

Over time, law enforcement agencies often lose officers due to a change in staffing or retirement. As agencies continue hiring new officers, they can replace the presence in the community, but there is a gap in knowledge and experience. Research suggests that using AI and ML in policing has already begun to bridge the gap, with facial recognition being one example. Inexperienced officers may make rash decisions, escalating a situation into a violent encounter, which might be avoided with the insight and experience AI could provide. 

Many officers are turning to jobs in the private sector, where there is potentially more pay and flexibility. Many opt for jobs where they do not have to interact with the public. The departure of these senior employees, though, leaves experience gaps. According to The Marshall Project, “The problems may seem more severe because the sharpest declines have been in cities with more than 1 million residents, such as New York, Los Angeles and Chicago, where the number of sworn officers dropped twice as fast as the national average.” 

This article discusses how creating a database of police experience to be accessed and used by newer officers can lessen the impact of losing senior and experienced officers. Using a mobile device, officers could access the expertise of the officers leaving the profession, resulting in better field performance by younger and newer officers. It would also slow the experience gap that is widening. 

The Exodus 

According to the U.S. Bureau of Labor Statistics, there was a steady decline in law enforcement and local government jobs during the pandemic. “From March 2020 to August 2022, the number of government workers dropped by 5%, while the number of local law enforcement employees decreased by 4%, the most recent data shows. The Census Bureau’s government payroll survey shows similar trends.” 

Turnover in a law enforcement agency is an important topic to review because replacing an officer can be time-consuming and expensive. Direct costs to the agency include background checks, uniforms and equipment, psychological assessments, medical assessments, overtime, training, and administrative costs. Indirect costs include quality of services, productivity, accumulated institutional, and a loss of professional knowledge and skills. CNN news reported shortages in 2022 of new police applicants and current officers resigning or retiring. Due to these factors and increasing community concerns, law enforcement agencies are predicted to increase police and detective jobs by 3% between 2021 and 2031, so stemming the flow of losses and reducing turnover should be a priority for the police. If cops began gathering the experience of veteran officers (many of whom are leaving), they could prevent the loss of institutional knowledge and use AI technology to address the most critical aspects of community confidence. 

A Catalyst 

In Minneapolis, Minnesota, in 2020, George Floyd was arrested in front of a grocery store and died as a result of the arrest by Minneapolis Police Department officers. Two of the arresting officers were employed for less than a week, with no prior police experience other than Basic Academy Training. Floyd’s death was international news, part of protests and legislation leading to nationwide changes in law enforcement policies and procedures. 

The community demands professional policing that is free of bias and relies on fairness and de-escalation tactics to offer alternative solutions. As a result, they may be relying on their limited training and peers for guidance. AI, though, could offer much more expertise to any contact in the field, which supports law enforcement efforts to implement AI and machine learning (ML) into the daily lives of officers on duty for better and safer communities. 

AI and the Police – Change the Approach 

In “Flash Foresight,” Daniel Burrus offers a different lens to examine the problem of losing senior staffing. All places of employment will face impactful losses as employees exit the agency. Burrus suggests that there are seven triggers for generating foresight, two of which are “transform” (i.e., using technology to build competitive advantage) and “go opposite.” Going opposite suggests going the opposite direction from where everyone else is looking, which results in seeing things nobody else is seeing – for example, solutions that no one could see because no one was looking for them or considering hidden opportunities, unnoticed resources and overlooked possibilities. 

Considering the loss of senior officers and the slow addition of newer officers, law enforcement can use technology to retain officers’ wisdom, allowing newer officers to become more efficient. To hire and retain officers, some agencies are already using AI systems. For example, the Austin Police Department uses Versaterm Public Safety’s Case Service to take non-emergency police reports. That AI software instantaneously communicates with the public through voice, mobile, web, and text messaging to ask questions and fill out reports like those typically provided by officers. 

The Austin Police Department is an example of going opposite and putting its energy into building up its AI system instead of trying to hire itself out of this situation. Researching AI technology, scientists and engineers at Pacific Northwest National Laboratory are working in the field of human-machine teaming to bridge the gap between today’s tools and the machine teammates of the future. The Digital Police Officer (D-PO) “is a vision of machine teammates: an artificial intelligence-based partner that can be reached through multiple devices including the patrol car’s on-board computer and officers’ mobile devices.” The interface between the officer and D-PO can be as simple as the officer’s smartwatch. 

How AI Will Be the Solution 

If an AI system were available to offer alternative solutions in the Floyd incident, it could have suggested alternative tactics, and the new officers’ actions may have changed the outcome. According to a Police Executive Research Forum Report, agencies in 2020-2021 typically filled only 93% of the available positions. Overall, responding agencies hired 5% fewer officers and saw an 18% increase in resignations and a 45% increase in retirements compared to the previous year. As these trends continue, more police departments will employ inexperienced officers. Inexperience may increase the potential for excessive-force issues, “No matter how well they are trained or what level of professionalism is instilled in them, many new officers lack the maturity and in groups may do things that are more emotional than wise.” 

To understand the future of AI, look at how law enforcement and others have implemented it in the past. John McCarthy defined AI in the mid-1950s as “The science and engineering of making intelligent machines.” During that period, AI research focused on creating machines that could independently perceive and respond to their environment and perform tasks that would typically require human intelligence and decision-making without human intervention. AI has often replaced human tasks, and it is now widely accepted to have AI running many industries, such as the auto industry assembly lines, and replacing the human worker. Those AI machines weld parts and paint the exteriors of cars at a level of precision that humans cannot match. In this example, the intricate detail the vehicle assembly requires could directly translate into machines performing the complex functions of police officers. An example would be giving officers step-by-step guidance on what questions to ask and what steps to take on an investigation. Additionally, AI can go through hours of video and audio evidence much faster than an officer can and determine the next leads to investigate or point the officer toward valuable suspect information. 

An AI system would be like “Siri” on an iPhone in these evidence searches. The system would use machine learning to guide the officer, like partnering with a 25-year veteran officer. Currently, AI systems are used for booking flights, monitoring travelers in airports, and assisting pilots flying airplanes. This technology can take a 911 call, find the location on closed-circuit television surveillance, watch the real-time actions of the subjects involved, and drive or navigate the officer to the scene. AI that learns from all the available video footage, radio traffic, and court outcomes could use that database to guide a newer officer to make decisions similar to a seasoned officer by suggesting senior-level strategies for handling that type of call. Pacific Northwest National Laboratory believes that, in time, D-POs will be deployed in law enforcement agencies. The outcomes can be remarkably different when it and other similar technologies enter the field. 

How the New “Digital Police Officer” Would Help New Officers 

A wearable, field-deployed AI for an officer will capture all the experiences, strategies, and techniques senior officers use to navigate citizen or internal interactions. Through each occasion, the AI system will learn positive or negative outcomes and then deploy guidance to field personnel that veteran officers provided in the past. For example, AI now assists pilots while flying planes; a mobile AI would assist officers on calls by giving them forecasted challenges. 

Using AI technology to identify where shots may have been fired so officers can respond more efficiently and effectively to shooting calls. Officers can respond to areas where current trends suggest a specific type of crime may occur (e.g., a burglary and the most likely times for that to happen). An officer can then be on the scene at or around that time and may deter or catch a suspect. Knowing where a crime is happening is a skill that takes experience in the field. AI-powered predictive policing could direct new officers to areas they would otherwise know only with the experience of responding to numerous calls and crimes in that specific area. AI takes the need for personal expertise out and uses the data to suggest where the officer should be to prevent crime. However, AI technology does not come without risk of bias becoming a factor. 

AI may provide the officer with real-time camera information that would guide new officers by directing who and where involved parties are and if there are any weapons. The interface could be seamless, like having a veteran officer working right next to the new officer and giving guidance. 

In the near future, a mobile AI will be deployed and offer the officer and public more confidence in the information to be processed and presented before deciding. AI will provide a superior component to policing because it removes the emotional or human element from a decision, which is heightened during an incident. It will remove any “caught up in the moment” or “I got pissed” feelings of an officer, which could potentially impact the decision-making. Like in every community, this research could also affect those who engage in criminal activity and would be aware that officers have more accurate information. AI could lower all involved parties’ stress, creating a calmer community. At a time when communities are looking at ways to re-imagine policing, AI and ML research could fulfill that expectation. 

The Challenges 

The introduction of AI was only 70 years ago; the next 70 years could see AI systems replacing officers entirely to better society. With all the advantages that mobile systems would offer, there are potential conflicts to consider when relying on AI in community settings. In previous studies, AI was utilized in differing applications, and it has exhibited decision-making bias, which would be a crucial factor in policing. For example, AI systems have shown bias in credit allocations in the banking industry and shades of racial bias in criminal sentencing trials. The community and agencies alike will want to know how the technology could reduce bias implications and potential ethical issues. 

AI will be evaluated on the testing outcomes and real-world examples. Research may also reveal a history of fairness. However, if the AI predicts incorrectly and an innocent person is unnecessarily affected, there is little accountability. The public is not fully exposed to the capabilities of AI yet, so there needs to be a greater understanding of how a decision would help create validity in the next decade. As they do, and as the police become more effective with far fewer adverse actions and decisions using AI, what now seems like science fiction will become the norm. 

One path to solving the problem of losing experienced law enforcement officers is using AI in the community. With the high rate of advancement and development coupled with the potential need, AI could be the most sought-after solution in law enforcement. AI fills the gaps for non-biased policing and is becoming a more widely accepted standard in law enforcement. Law enforcement leadership should be prepared for their organizations, communities, budgets, and legal teams to integrate and adapt to that future reality. As a result, AI partners could make the current and new generation of officers more efficient, safer, and able to provide a high-quality law enforcement service from the first day an officer starts.