Montreal Unveils New AI Tool to Prevent Metro Suicides

Currently, the AI system accurately identifies one out of four people who are likely to attempt suicide.
Representative image
Representative imagewirestock

AI’s true potential lies in impacting human lives at the grassroots level and that is what Montreal's public transit authority, the Société de Transport de Montréal (STM), is doing, in collaboration with the Center for Suicide Intervention (CRISE). The team is developing an AI system for suicide prevention in the city’s metro stations. The pilot project involves utilizing CCTV footage from the stations to identify potential distress signals in individuals.

Brian Mishara, Director of CRISE and Co-investigator of the STM AI initiative, explained that the project analyzes videos of individuals who have attempted suicide in the Metro. The goal is to identify behavioral indicators that may suggest someone is in distress. Due to the impracticality of human monitoring of numerous screens, AI plays a crucial role in recognizing these signs.

"We've got some indicators, but obviously a human cannot watch hundreds of screens all day long to try to identify those behaviors," Mishara added.

How does this work?

The AI system, devoid of facial recognition software, identifies distress indicators, prompting immediate interventions to prevent self-harm. The STM predicts real-time warnings to the control room or directly to metro operators, enabling proactive measures such as train braking and the deployment of special constables to the location.

Currently, the AI system accurately identifies one out of four people who are likely to attempt suicide. To further enhance safety, the possibility of installing barriers to prevent platform jumps has been considered. However, this solution is deemed expensive.

The STM expressed optimism about the AI pilot project, deeming it "promising." It aims to implement the system within two years and reinstate the goal of introducing platform screen doors as part of the 2023-2033 plan, allocating $5 million for project evaluation.

Related Stories

No stories found.
CDO Magazine