Montreal’s police force has recently enhanced its surveillance capabilities by integrating artificial intelligence (AI) into its video monitoring systems. This technological upgrade, initiated by the Service de police de Montréal (SPVM), began during the summer months and aims to provide real-time video analysis from numerous city cameras.
Implementation of AI in Surveillance
The SPVM acquired a robust video analysis software incorporating AI to streamline investigations. This system is capable of analyzing up to 120 live video streams concurrently and can support a minimum of 700 users simultaneously. This advancement is expected to significantly improve efficiency and response times for police investigators reviewing video evidence post-crimes.
Real-Time Analysis Capabilities
The AI-powered surveillance enables real-time identification of vehicles by their license plates and individuals based on their clothing. Such technology capitalizes on the functionalities provided by specialized video analysis software like BriefCam, known for isolating pedestrians by their appearance. Authorities emphasize that while real-time analysis is a potential tool, it will primarily be used in extreme emergencies, such as active shooter situations.
Concerns Over Privacy and Scope of Use
Despite its intended limited use, experts like Tamir Israel from the Canadian Civil Liberties Association (ACLC) express concern about the possible expansion of AI capabilities beyond initial plans. Israel warns that once integrated with city networks, additional functions may be activated without public disclosure, potentially including facial recognition technology.
- Current camera count: 46 in the city.
- Contract value for AI software: $1.8 million over five years.
- Seller: iMotion Sécurité, a Quebec-based provider.
Public Transparency and Software Selection
There has been significant controversy surrounding the AI video analysis software. Various regions in Canada have adopted technologies from companies like Idemia and BriefCam, both offering different capabilities but also having faced scrutiny for prior improper use. Unlike other police departments that host informational sessions on algorithmic bias and privacy, the SPVM remains tight-lipped about the specific software being utilized.
This lack of transparency has drawn criticism from civil liberties advocates, who argue that understanding the tool’s specifications is essential for public trust and informed discourse on surveillance practices. As cities like Montreal navigate the balance between security and privacy, the broader societal implications of AI surveillance remain a critical debate.
