AVA SEND HELP!
BY DON REDMOND | FBINA SESSION #263
“Ava! Send Help!”
The 9-1-1 call was frightening. The female caller was screaming into the phone that her boyfriend had just been stabbed at the marina by two suspects leaving in a white vehicle. The police dispatcher immediately identified the three closest police units by using GPS coordinates from the caller’s phone and dispatched the fire department, ambulance, and launched a surveillance drone from the roof of the police department.
Additionally, video cameras in the area were activated, capturing the suspects’ vehicle license plate and pictures of the vehicle and suspects, which were then sent to the responding officer’s mobile computer. The dispatcher calculated the probable direction of travel and officers soon intercepted the fleeing vehicle. Within minutes of receiving the frantic 9-1-1 call, police had the suspects in custody. Once again, Dispatcher Ava was credited with providing officers with all the information they needed to solve another violent crime. Dispatcher Ava is not human, though – “she” is the Autonomous Voice Activation system (AVA).
Today Shaping Tomorrow
The City of Chula Vista is the second largest city in San Diego County, encompassing 50 square miles with a population of 267,842 residents. The Police Department employs 246 sworn police officers and 90 professional support staff. The Chula Vista Police Department (CVPD) has a reputation of being a progressive agency that prides itself on being cutting edge with technology and proactive in seeking solutions.
Like almost every other law enforcement agency of a similar size, Chula Vista also staffs a full-service dispatch center. Although the center’s functions are essential to the delivery of police services, they are expensive, difficult to staff and unable to control the vast amount of data coming into and going out of the communications center. For instance, CVPD employs 25 dispatchers and anticipates hiring another seven over the next two years. Each dispatcher costs about $70,000 each year, and the center has a total budget of more than $250,000. Finding qualified applicants who can answer 9-1-1 phone calls and multi-task under extreme pressure is challenging. Additionally, it takes six to nine months to train a dispatcher to be competent enough to work on their own. As is common in the profession, for every three dispatchers hired, only two will be successful in training. With the continual rise in emergency call volume, including text messaging, pictures and videos, the Department will need to hire more dispatchers to handle the increasing workload unless other means of doing the job can be found.
What if, however, a police department could eliminate police dispatchers from the annual budget? Could virtualization and automation of the communications center remove the need for humans to perform the dispatch function altogether? In fact, there are places where this is already happening. In Dubai, virtual dispatch centers exist that answer non-emergency calls and can walk citizens through the process of filing a police report. In Copenhagen Denmark, AI-powered technology called Corti is being used to augment emergency dispatchers in EMS communications centers. The Journal of Emergency Medical Services states, “Corti quickly assists dispatchers in concluding what is happening by finding patterns in the caller’s description. Corti analyzes the full spectrum of the audio signal, including acoustic signal, symptom descriptions, tone and sentiment of the caller, as well as background noises and voice biomarkers. These distinctive features of the call are immediately and automatically sent through multiple layers of artificial neural networks that look for patterns that might be useful for the dispatcher.” Could this technology be expanded to perform the full array of duties as a virtual police dispatcher? It is useful to explore a possible future by traveling into a police department where they are working to integrate AVA into the fabric of their organization.
A Glimpse into the Future…A Scenario
It is a bright fall day in 2022. As has been the case for years, the River City Police Department (RCPD) is unable to find quality dispatch applicants and is faced with a financial crisis relating to the exorbitant cost of the 9-1-1 Communications Center. The Department contacts Future Insight, a technology firm specializing in artificial intelligence (AI) and cognitive analytics to test whether AI can be used to supplement or replace police dispatchers to fulfill the mandates of running an emergency call center.
AI being used to answer phone calls is not a new concept. Rarely do companies have employees answering telephones. For example, Google announced they have partnered with companies like Cisco and Vonage to actively focus on using AI to replace call center workers. Google Cloud Chief Scientist Fei-Fei Li states, “When we studied the challenges faced by real contact centers every day, we found that customers often have simple transactional or informational requests.” Google’s machine learning-powered customer representative using Cloud speech-to-text for accurate speech recognition answers the customer’s call. The customer is immediately greeted by a Virtual Agent that answers questions and fulfills tasks all on its own. If the customer’s needs surpass the Virtual Agent, the caller is transitioned to a human representative.
The public probably will not care that it is a computerized voice answering their 9-1-1 call. They just want a cop to show up fast. Most people may forget they are speaking to a computer since they have come to accept the integration of technology into everyday life. Recognizing an opportunity, RCPD officials meet with researchers from Future Insight to discuss the possibility of using their technology in police dispatch centers. But one issue that RCPD was not prepared to answer was how “real” artificial intelligence should become.
AI has been developed that learns and make decisions, but should it also have emotional intelligence, empathy, and even doubt itself? Developers point out that Siri, Alexa, and Cortana are not judgmental or reactive in attitude. They are machines that do what they were told. Would the community rather have AVA unemotional and lacking in empathy, or should she question a caller’s motive and even argue? Officials ultimately decide AVA should be understanding but detached of emotions, thinking the world is not ready for computers being capable of producing emotions like resentment, blame, gratitude, guilt, indignation, or pride. The fear is that if AVA has emotions, she could become angry at callers, which could impact how she decides to send help or delays her response.
RCPD did not approach their decision to automate their dispatch center lightly. They learned a lot from AI-technology leader IBM, which led the development of “Watson,” their ever-expanding AI platform. Watson’s goal is “to tackle increasingly difficult real-world problems.” In this effort, Watson continues to gain learning capabilities enabling decision making and allowing for greater economic and societal benefits according to IBM’s senior VP for Cognitive Solutions and Research John Kelly III. Additionally, RCPD studied the report released by the federal government in 2016 identifying the potential impacts of implementing AI technology to mitigate costs including the “positive contributions that will aggregate productivity growth.” They also visited the City of Dubai, which became the world's first unstaffed "smart" police station, allowing citizens to pay parking tickets, track criminal investigations, and report crimes to real police officers using video calls and saw firsthand the success of using AI-technology to increase efficiency. As a result of their work to develop a suitable system, RCPD announced that, by the end of the year, River City would become the first law enforcement agency in the country to have an autonomous dispatch center.
RCPD and Future Insight recognized they would be setting a new precedent for law enforcement. Although the software engineers caution the growing AI technology needs more time to advance, the partnership between big business and local government is an opportunity no one wants to oppose. Future Insight and city attorneys created a contract for the company to provide the technology and engineers to set-up and run the dispatch center, and RCPD agreed to use the money saved from reducing staff towards the cost and maintenance of the equipment.
On January 1, 2023, AVA went live, using AI and cognitive analytics to answer 9-1-1 calls and dispatch officers for service. The transition from human to virtual dispatcher was encouraging. As expected, most citizens calling the communications center never realized they were speaking to a computer. AVA assisted callers and was extremely effective in dispatching resources. But there were some instances where AVA did not understand what was needed during an emergency. On occasion, tow trucks were sent instead of ambulances, police officers were dispatched to the wrong addresses, and poor connectivity between the caller and AVA jeopardized the concept of an autonomous dispatch center.
It would take over a year for engineers to overcome all of the issues and problems. During that time, AVA began learning from her mistakes and even anticipating the needs of callers and officers. AVA now monitors the city and officer body-worn cameras with facial recognition and warns officers of dangerous or wanted individuals. As a result of AVA’s efficiencies, police response times and crime statistics are reduced, community members and officers feel safer, and other police agencies begin inquiring how they can implement the technology.
By 2026, law enforcement and community members have recognized and accepted the positive effects of autonomous dispatching. AVA is now a “member of staff” providing for the safety of officers and the community. By 2030, many police agencies follow in RCPD’s footsteps. Just a decade earlier, law enforcement was working in the old ways. Now, AVA has cut costs, reduced inefficiencies, and enhanced service to the public. Newer employees wondered how they had ever done it as it had been done in the past.
Preparing for Tomorrow
This scenario is a peek behind the curtain of how AI-technology and cognitive analytics can be used in autonomous dispatching. Law enforcement is already using elements of these technologies to track user’s electronics and predict an individual’s probability of committing an offense in the future. Police agencies are moving to more data-driven approaches in fighting and solving crimes and having the ability to analyze a 9-1-1 call immediately, access multiple databases, and provide real-time information to police officers that will transform the expectations of police communication centers.
Some will argue that communities will never accept speaking to an AI dispatcher. The future, though, will look very different than today. Scientists are developing computers that learn and make decisions and may even become sympathetic and empathetic. Google has developed Google Duplex; an AI personal assistant featuring a human-sounding computer. As a result, Google already has a track record of real conversations between Duplex and people on the phone who did not even know they were talking to a robot.
With a reduction of workforce in police dispatch centers, will the public be inclined to interact with an automated computer instead of a real person, or will the gradual increase in daily interaction with AI personal assistants lead to an expectation that law enforcement will follow suit? Furthermore, will technology ever develop to the level of effectively replacing people in law enforcement communications center without compromising compassion and community trust? Law enforcement leaders must anticipate a future of autonomous dispatch centers and ensure any implementation has a positive effect on community relations and their police department.
Law enforcement communication centers have only been in existence for approximately seventy years. But in that short time, advancements in technology allows citizens to immediately call law enforcement for help and speak to a police dispatcher. Before the modern law enforcement dispatch centers, people would scream for help and officers would use bells or whistles to summon reinforcements. Our grandparents may remember a time from their childhood when they did not have a telephone, but somehow, they were always able to get help in their time of need. AI-driven technology may take us back to the time without police dispatchers. A future may exist when someone calls 9-1-1 and Dispatcher AVA answers. Policy makers and stakeholders must anticipate this future and ensure these changes have a positive effect on community relations and their police departments.
About the Author:
Lieutenant Don Redmond began his law enforcement career with the Chula Vista Police Department in 1996. Lt. Redmond has worked various assignments throughout his career including Patrol Officer, SWAT, Field Agent, Patrol Sergeant, Professional Standards Unit Sergeant, Watch Commander, Community Policing Commander and Professional Standards Unit Lieutenant. Currently, he is the Manager of the Police Department Jail and 911 Communications Center. Lt. Redmond has a Bachelor of Arts Degree in Education from San Diego State University and a Master of Arts Degree in Administrative Leadership from the University of Oklahoma. He is presently attending the California POST Command College (Class 64) and is a graduate of the FBI National Academy Session 263.
BACK TO MAGAZINE...