AI at 100 Montreal Intersections
By Maxime Bergeron, La Presse
Disorganized construction sites and streets overrun by orange cones? It’s a familiar refrain in Montreal.
This next verse, however, is more surprising: the city's main arteries are also packed with technology. More than 700 cameras monitor traffic at all times. Hundreds of sensors and counters are scattered across intersections. Traffic lights can be controlled remotely. The list goes on. This high-tech armada places Montreal in the global vanguard of so-called "smart cities."
Yet, the general impression on the streets is often chaotic. Retro, even. Just think of police officers being paid overtime to manage traffic by manually pressing a button...
How can we make better use of all these electronic gadgets? The City of Montreal thinks it has (finally) found a solution: artificial intelligence (AI).
The Martinez Ferrada administration will announce this Monday the installation of AI-based systems at 100 of the city's intersections by the end of the year. This marks the continuation of a project initiated at the end of Valérie Plante's administration.
Lessons from Australia
I learned this news upon returning from Australia, where I went to observe urban innovations in Melbourne and Sydney.
Both cities have successfully reduced congestion, improved safety for vulnerable road users, and achieved efficiency gains without breaking the bank. (You can read my findings in a series of reports published in La Presse).
How did they successfully make this shift?
First: By leveraging the resources they already possessed—like their networks of sensors and cameras.
The Missing Link: These devices were collecting data on the passage of cars, pedestrians, cyclists, trams, and so on. But they lacked a framework to make sense of this mountain of information.
That is where AI came in to change the game. The technology analyzes millions of movements across the road network in real time. As it "learns," the AI detects patterns. It can then propose better traffic light sequencing or more efficient detours around construction sites.
While all of this is recent, the Australian results are promising.
The Montreal Plan
Montreal hopes to achieve similar gains. The City has awarded a $2.5 million contract to the Quebec firm Tacel, which distributes technology from the American company DERQ.
To put it simply, the group will install small control boxes on traffic lights. These will connect existing cameras to an AI-powered system. This mechanism is already in use in several cities, including Detroit and Dubai.
Montreal has primarily paid for software licenses covering a 10-year period. There will be no outsourcing: the technology will be managed internally by employees at the Urban Mobility Management Center (Centre de gestion de la mobilité urbaine or CGMU).
The AI system has been quietly deployed at around thirty Montreal intersections over recent months. The City is aiming for one hundred by the end of 2026. This would make it one of the largest networks of its kind in North America.
Healthy Skepticism and Privacy Concerns
The potential of this initiative is as high as the disappointments of the past.
Between the purchase of 100 "smart" cameras under Gérald Tremblay, the creation of a "Smart City Office" under Denis Coderre, and the launch of a "Mobility Squad" under Valérie Plante, Montrealers have been promised a mobility revolution far too often. They have every right to be skeptical...
Privacy concerns will also arise. Millions of images captured on Montreal streets, analyzed by American software? The City will need to demonstrate that it has put up all the necessary safeguards.
"We do not depend on any foreign supplier. Our data stays with us. It goes into our servers, in one of our buildings downtown," Alexandre Teodoresco, the executive committee member responsible for optimization, municipal performance, and innovation, told me in an interview.
Looking Ahead: A "Digital Twin"
Despite legitimate fears, Montreal intends to move further and faster in adopting new technologies—with AI leading the charge.
The mayor has committed to creating a "digital twin" of the metropolis, similar to those that exist in Singapore, Rotterdam, Sydney, and many other cities. Essentially, this involves creating a digital replica of the city to run various simulations—for example, mapping the impact of a specific construction site on traffic congestion. The more precise data it is fed, the more useful this digital twin will prove to be.
The Martinez Ferrada administration also wants to transform downtown into an experimental zone—a sort of living lab. The idea is to bring local tech SMEs and the university research sector together to find better ways to plan construction projects.
The "brains" of an entire ecosystem will thus be put to work. This bodes quite well, as Montreal is certainly not lacking in that department.