8.17.23 – Officer
Technology Contributing Editor Toni Rogers discusses the strengths and necessity of artificial intelligence in emergency communications centers.
Technology continues to grow rapidly in all industries, including policing. Facial recognition, license plate readers, predictive policing, gunshot detection systems, early warning systems in training databases, biometrics, and so many others are used regularly by many police agencies. Artificial intelligence (AI) is finding a home in 9-1-1 and public safety communications centers.
The most basic function of AI is to respond independently within a specified environment to complete tasks usually completed by a human. What makes AI different than previous programs running a set of commands is that AI can learn from its own experience and make some adjustments. For example, in fraud detection, AI learns from valid alerts to improve the quality of fraud notices generated in the future. That same set of learning is applied to facial recognition used in policing, as programs learn how to find faces more quickly and accurately.
AI answering the call
In Portland, Oregon, some of the city’s nonemergency telephones are being answered by an automated attendant using AI. Working through system testing, callers to the 311 number would work through the automated attendant to get their questions answered or be transferred to the appropriate city department. Like other AI telephone attendants in corporate customer service, the attendant answering can transfer callers and text information to callers providing website addresses or other information.
When reviewing the possible use of AI, the study commissioned by the City of Portland determined that using AI to answer nonemergency calls could reduce the Bureau of Emergency Communications call volume by almost 17%. One of the City’s concerns is that the growing call volume in Portland has made it difficult for the communications center to meet the national standard for answering 9-1-1 calls. Using AI for nonemergency call response frees call-takers to answer 9-1-1 calls, providing more timely assistance and shorter response times.
Real-time language translation
Language translation in 9-1-1 and communications centers traditionally worked through a third-party translation service. While the process can sometimes be time-consuming and challenging for the caller and call-taker, it was the best option. With AI, the Presidio County Sheriff’s Office in Texas now uses live audio language translation for Spanish. The AI integration translates what the caller says into real-time text on the call-takers computer screen.
Before AI, a Spanish-speaking caller would be transferred to a translation service, a code entered, and then would begin speaking to repeat what the call-taker asked and repeat the answer in a back-and-forth translation process. AI saves time, and callers are likelier to stay on the line and ask for help. The average time on an English-language 9-1-1 call is about 2 minutes, yet the average for a non-English-language 9-1-1 call is approximately 6 minutes. The translation AI can reduce the time substantially, reducing response times, freeing up call-takers for other 9-1-1 calls, and capturing the translated text for the database.
Tracking stressful calls
Emergency communication centers are routinely understaffed. Less staff equates to more calls handled by a dispatcher and less time for breaks and time off. Compassion fatigue is associated with jobs that routinely try to help others who share information about one of the worst times in their lives. Working in a position to keep both the officers and the public safe, call takers need to be in the best frame of mind, and often, they will not let anyone know if they need a stress break.
Using AI, North Central Texas Emergency Communications District (NCT911) tracks dispatchers’ calls to identify call takers who need a break. The program will then alert a supervisor, who can check with the dispatcher directly. NCT911 is testing a program to inform them when dispatchers have numerous bad or tragic calls. Using keywords and the caller’s stress using voice tones, the program monitors the number of stressful calls for each dispatcher and lets the supervisor know when a preset threshold is reached.
AI is here to stay
Artificial intelligence continues to improve, and police agencies are already using AI to provide real-time data analysis and investigative assistance. The future will see communications centers manage video, images, text, and information from various sources, and some agencies are already using AI in their communications centers. There is room to incorporate AI in centers that do not yet use the technology to assist call takers and dispatchers.
Technology is improving, and so are the AI programs that will be helpful in emergency dispatch. Being open to AI and technology’s possible solutions in the communications center can help bridge service gaps due to staffing and other limitations. And at the end of the day, communications center improvements through technology and AI can help keep police officers and the public safe.