Toronto Police Deploy AI to Handle Non-Emergency 911 Calls
Table of Contents
Toronto police are now using artificial intelligence (AI) to screen 911 calls and direct those deemed non-emergency to a specialized agent, aiming to reduce wait times for critical incidents. Teh system, developed by Hyper, was launched on January 22, 2024, and is designed to handle routine calls like noise complaints or requests for facts.[https://www.cbc.ca/news/toronto/toronto-police-911-ai-1.7084489](https://www.cbc.ca/news/toronto/toronto-police-911-ai-1.7084489)
How the AI System Works
The AI system engages callers in a conversational manner to determine the nature of their emergency. If the system identifies an emergency or the situation escalates during the call, it instantly transfers the caller to a human 911 operator. According to Toronto Police service spokesperson Stephanie Watts, the goal is to free up human operators to focus on urgent situations. [https://www.cbc.ca/news/toronto/toronto-police-911-ai-1.7084489](https://www.cbc.ca/news/toronto/toronto-police-911-ai-1.7084489)
The Halton Regional Police Service (HRPS) has been utilizing similar technology since November 2023, reporting a decrease in 911 wait times consequently. [https://www.cbc.ca/news/toronto/toronto-police-911-ai-1.7084489](https://www.cbc.ca/news/toronto/toronto-police-911-ai-1.7084489) According to halton police Const. Jeff Dillon, the implementation has been prosperous in reducing delays.
Previous AI Use by Toronto Police
This isn’t Toronto police’s first foray into AI. In October 2019, the service admitted to using the controversial facial recognition tool Clearview AI, but committed to discontinuing its use in Febuary 2020. [https://www.cbc.ca/news/toronto/toronto-police-911-ai-1.7084489](https://www.cbc.ca/news/toronto/toronto-police-911-ai-1.7084489)
Addressing Public Concerns
Watts emphasized that a human operator will always be available to answer 911 calls, even with the introduction of AI. However, the use of AI in such a critical public service raises concerns for some experts.
Morteza Zihayat, an assistant professor of human-centred AI at Toronto Metropolitan University, questioned how the system would handle calls that don’t clearly fit into the non-emergency category. He highlighted the potential for misclassification and the challenges of escalating situations.”What if there is a misclassification and your systems do not understand the context in the way that we as humans understand? Or maybe it’s a non-emergency situation, but it’s escalated quickly to an emergency situation. So what’s going to be the process there?” he asked. [https://www.cbc.ca/news/toronto/toronto-police-911-ai-1.7084489](https://www.cbc.ca/news/toronto/toronto-police-911-ai-1.7084489)
Hyper CEO Ben Sanders explained that the system is designed to redirect calls to a human operator if an emergency is detected or a situation escalates. He also noted that this is an advancement over the previous system, where callers mistakenly reaching the non-emergency line faced longer wait times. [https://www.cbc.ca/news/toronto/toronto-police-911-ai-1.7084489](https://www.cbc.ca/news/toronto/toronto-police-911-ai-1.7084489)
Zihayat also expressed concerns about the potential for the AI to provide inaccurate information, such as oversimplified police procedures or incorrect reporting instructions. He cited examples like ChatGPT, which can sometimes confidently provide incorrect or incomplete information. [https://www.cbc.ca/news/toronto/toronto-police-911-ai-1.7084489](https://www.cbc.ca/news/toronto/toronto-police-911-ai-1.7084489)
Hyper’s Response to Accuracy Concerns
Sanders countered these concerns, stating that Hyper operates within “strict guardrails.” He explained that the system only provides answers based on agency-approved data and will transfer the call to a human operator if it cannot provide a reliable answer, avoiding the risk of misinformation. [https://www.cbc.ca/news/toronto/toronto-police-911-ai-1.7084489](https://www.cbc.ca/news/toronto/toronto-police-911-ai-1.7084489)
Privacy Considerations
Zihayat also raised questions about caller privacy. Sanders assured that Hyper has prioritized data security, implementing strong security controls, limited access, and strict use policies aligned with government and law enforcement requirements. He confirmed that caller data is owned by the police department and is not used by Hyper for training or any external purpose. [https://www.cbc.ca/news/toronto/toronto-police-911-ai-1.7084489](https://www.cbc.ca/news/toronto/toronto-police-911-ai-1.7084489)
Key Takeaways
- Toronto police are using AI to screen 911 calls and direct non-emergencies to specialized agents.
- The system aims to reduce wait times for critical incidents by freeing up human operators.
- Human operators will still answer all 911 calls, and the AI will transfer calls if an emergency is detected.
- Concerns have been raised about potential misclassification, inaccurate information, and caller privacy, which Hyper is addressing through safeguards.
Worth a look