Advertisement

Your City

Halton police is now using AI to pick up non-emergency calls, but one expert says the pilot might cause frustration

Halton police turn to AI for non-emergency calls, a move some say could improve efficiency but risk alienating residents.

Multiple women working at a busy emergency call centre with multiple computer screens and communication equipment in an indoor office setting.
Halton police announced the official launch of SARA, an AI system that will help police handle certain non-emergency calls. (Courtesy: Halton Regional Police Association)

What to know

  • Halton Regional Police Services launches SARA, an AI system to manage non-emergency calls and ease staff workload.
  • The tool can redirect calls or send helpful links via text or email.
  • Expert warns the system may cause frustration and make callers feel unheard.
  • The launch reflects a growing trend of AI use in government services amid public concern.

Halton Regional Police Service (HRPS) hopes that their newly-launched Artificial Intelligence (AI) system to handle its non-emergency calls will enhance safety in the long run, but one expert says it could also cause public frustration. 

On Monday, HRPS announced the official launch of SARA (Smart Answering Routing Assistant), an AI system that will help police handle certain non-emergency calls.

Starting this week, those calling the HRPS non-emergency number are now greeted by SARA and asked a series of short questions, as the system determines if the call could be handled automatically or transferred to a human communicator. 

According to police, calls will be automatically transferred to communicators if: 

  • There’s an emergency; 
  • The system isn’t able to understand the caller; 
  • The caller needs translation services; 
  • The system detects unusual activities or discrepancies;
  • The system determines it can’t help the caller. 

Advertisement

HRPS tells Now Toronto that besides redirecting calls, the system is also able to email and text links to help callers find online report forms, resources and information for other services not provided by police. 

The system was developed by the police service’s Communications Bureau and Information Technology Unit in collaboration with tech company Hyper

Earlier this year, the company had announced a $6.3-million fund to help 911 services across the U.S. implement AI systems to pick up calls, which they say would expedite responses and help dispatchers focus on more important tasks. 

According to police, the SARA system has been “tested extensively earlier this year” before its launch date, with HRPS Deputy Chief Jeff Hill sharing optimism for the new tool. 

“For several months, our Communications Bureau and Information Technology Unit has been working in close consultation with the Hyper team to develop, train, and refine SARA, repeatedly testing a wide range of scenarios to confirm it’s able to receive and handle calls appropriately,” Hill said in a statement on Monday. 

“We’re not only confident in SARA’s performance, but we’re proud to be adopting such reliable and cutting-edge technology that will help us reduce the number of non-emergency calls our Communicators spend time on, which at the moment is approximately 160,000 per year.”

Advertisement

EXPERT RAISES QUESTIONS

University of Toronto’s Human-Centred Data Science Professor Shion Guha tells Now Toronto that for him the HRPS’s announcement raised a series of unanswered questions about the functioning of the program, public consultation about its implementation, and results of its testing period. 

The expert explains that the use of AI in governmental services is usually implemented to reduce costs or improve the quality of services provided. However, there is no evidence that proves these outcomes are indeed possible. 

In the case of police services, Guha says that problems with the system might emerge as the AI might have a different idea of which non-emergency issues are considered priorities and require quicker attention. 

“Imagine…you’re trying to get the AI to be like, ‘Hey, a tree has fallen on my property, and I need 311 to get out there and…get rid of it. That’s a good use case for a non-emergency call. What happens if the AI totally understands that a tree has fallen on somebody’s property, but instead it says, ‘Oh, I understand the tree has fallen on your property, but now you are deemed as low risk, so we will not deal with you right now.’ What happens then?” he questioned. 

For that reason, the expert believes these services have the potential to make callers feel like their problems are not being heard. Meanwhile, the expert says the system may also cause frustration if SARA isn’t capable of correctly understanding what the caller is trying to say. 

Advertisement

“When you are calling the police department, even on the non-emergency line. You have a human problem that you need the government to resolve. You have a human problem. And if people feel like there’s a distance between the government and their problems, then that increases a lot of frustration,” he added. 

WILL THIS BECOME A TREND? 

Implementing AI in governmental processes seems to be a growing trend across the world, including Canada. 

With Prime Minister Mark Carney’s first budget allocating billions of dollars for boosting day-to-day operations with AI and other non-emergency call assistance projects being launched across the world, Guha says he would not be surprised to see more police divisions adopting the new system. 

Read More

The expert says he has studied similar systems in cities like Pittsburgh and Wisconsin, which were implemented years ago, but the programs were met with significant criticism from the public. 

“The Carney administration, they seem to be thinking about AI and government automation as a thing that will help us in these upcoming times of austerity and stress that apparently we are going down. So, it’s a thing that’s happening. The [issue] is: the public is not ready.” 

Advertisement

On the other hand, Guha says there is opportunity to expand these AI programs if governmental departments, such as the police, are willing to collaborate with experts for the development of more effective systems. 

“Instead of thinking about all of these comments as a critique, …we should take it as an invitation and as an opportunity to collaborate together and see what we can do to benefit the public.” 

Advertisement

Exclusive content and events straight to your inbox

Subscribe to our Newsletter

This field is for validation purposes and should be left unchanged.

By signing up, I agree to receive emails from Now Toronto and to the Privacy Policy and Terms & Conditions.

Recently Posted