Montreal police (SPVM) are deploying AI-driven surveillance tools to monitor vehicles and individuals across the city. the use of high-powered cameras and video analysis software has sparked concerns regarding transparency and the potential for discriminatory policing.
A $17,134 subscription for Cloudrunner units
The Montreal police service (SPVM) has begun integrating advanced AI tools into its surveillance network, a move supported by a $17,134 contract awarded to Genetec last October. This contract secured four Cloudrunner units on a one-year subscription to assist in law enforcement efforts. Unlike the Flock camera systems frequently used in the United States, which share data across a central database accessible to multiple police forces, the Genetec system is managed by its specific operator, meaning the data remains under the control of the local agency.
Confusion over the Henri-Bourassa Boulevard sighting
A Radio-Canada reporter recently identified a high-powered camera mounted on a lamp post along Henri-Bourassa Boulevard in Montreal's Rivière-des-Prairies neighbourhood. This sighting led to a moment of public uncertainty when SPVM Chief Fady Dagher initially claimed he had no knowledge of the device on a Wednesday morning. As reported by CBC News, Dagher later clarified that the device was indeed an SPVM tool intended to combat vehicle thefts and armed violence, attributing his earlier confusion to distractions regarding a separate investigation into alleged racism at a Montreal-Nord station.
SPVM spokesperson Marc-Antoine Bélanger added that the specific camera was removed to protect ongoing investigative strategies. This removal, while standard for tactical operations,leaves the public with little visibility into how these high-powered devices are being deployed in residential areas.
The SPVM’s 2023 admission of civil liberties risks
Internal documents suggest that the police department is aware of the potential downsides of this technological expansion.. According to an internal privacy impact assessment obtained by CBC News, the SPVM acknowledges that these tools pose significant risks to civil liberties. The assessment, prepared in 2023, specifically flags the danger of placing innocent bystanders under investigation and the possibility of discrimination based on visible social characteristics.
To mitigate these risks, the department claims the software is configured to search for specific objects—such as a "red coat" or a "green bike"—rather than identifiable human traits or ethnicity. This "object-based" approach is intended to provide a buffer between automated surveillance and the profiling of individuals.
The unverified boundary between object tracking and biometrics
While the SPVM claims to be using AI for object recognition, significant questions remain regarding the true extent of their surveillance capabilities. SPVM spokesperson Anik de Repentigny denied that the force is utilizing facial recognition or biometric features of the software, stating instead that the tools help investigators sift through massive amounts of video. However, the scope of the data being analyzed is vast, covering the SPVM's 46 marked urban security cameras as well as private residential and commercial footage.
This private video is manually collected by officers, often with the consent of citizens or through a warrant.. The central unanswered question is whether the "object-based" search parameters are strictly enforced, or if the capability for biometric tracking remains a dormant risk within the Genetec system. Without more rigorous third-party auditing, the line between tracking a "red coat" and tracking a specific person remains dangerously thin.
Comments 0