Anyone who has lived in a mid-sized North American city will instantly recognize the experience of standing at a busy intersection in Kirkland, Quebec, on a Tuesday morning during rush hour: the slow accumulation of cars, the light cycling through its preset sequence regardless of whether thirty cars are waiting on one side and two on the other, and the low-grade frustration of watching seconds tick by while the system functions as if the traffic patterns of 1995 still apply. That experience has been shifting in some areas of Kirkland. In the corridors where they have been installed, AI systems that modify signal timings in real time in response to actual vehicle counts and movement rather than predetermined schedules have significantly shortened rush hour travel times. On its own, the difference isn’t significant enough to make news, but when you add it up over millions of commutes and thousands of junctions, it begins to seem like something to be concerned about.
Cities in Canada have been gathering justifications for experimenting. In many metropolitan corridors, the return-to-office patterns that emerged following the epidemic years brought vehicle numbers back to pre-2020 levels, frequently to infrastructure that had not been significantly improved in the meantime. Traditional pre-timed traffic lights were simply not built to handle the bottleneck caused by rising car ownership, changing commuting patterns, and the overall unpredictability of urban activity. When such lights were designed, traffic engineers could quite accurately anticipate peak times and adjust timings. The lines that back up through junctions every evening are one way that the city that actually exists in 2026 differs from those models.
| Category | Details |
|---|---|
| Topic | AI-Driven Traffic Management in Canadian Cities |
| Key Cities | Kirkland (QC), Quebec City, Calgary, Montreal, Ontario municipalities |
| Technologies Used | AI signal synchronization, smart cameras, radar, NoTraffic platform |
| Key Partnership | Quebec City + Google Green Light project |
| Primary Goals | Reduce congestion, lower emissions, improve intersection safety |
| Calgary Focus | AI “near miss” analysis at hazardous intersections |
| Kirkland Result | Significant rush hour travel time reductions |
| Environmental Benefit | Reduced stop-and-go driving = lower GHG emissions |
| Safety Technology | Real-time detection of vehicles, cyclists, pedestrians |
| Research Partner | Innovation UBC (camera/radar infrastructure) |
| Reference Website |
The infrastructural argument gains additional importance from the environmental aspect. Stop-and-go driving produces disproportionately high emissions in relation to the distance traveled because of the frequent accelerating and braking that define crowded urban traffic. Without needing any modifications to the vehicles utilizing the road or the conduct of the drivers inside them, an AI-optimized signal that minimizes the number of needless stops across a major arterial route serves as both a congestion management tool and a greenhouse gas reduction measure. Because it enhances current infrastructure without requiring physical development, Quebec City’s collaboration with Google’s Green Light project—which use AI to monitor traffic patterns and suggest signal timing adjustments—represents an approach that appeals to city planners. The hardware is already present. The work is done by the software.
The third and perhaps most persuasive point for the general population is safety. NoTraffic, a platform being evaluated in Canadian applications, tracks activity and anticipates conflict sites before they become collisions by using cameras and sensors at junctions to detect cars, bicycles, and pedestrians in real time. When a car turns right, the conventional traffic signal is unaware of whether a cyclist is approaching in a bike lane. In order to lower the likelihood of contact, an AI-enabled intersection can register both at the same time and perform the necessary micro-timing adjustments. Calgary has expanded on this reasoning by employing artificial intelligence (AI) to perform what the industry refers to as “near miss” analysis, which identifies intersections where close calls occur unusually frequently, even in the absence of reported accidents. Infrastructure investment may be focused on areas where risk is truly concentrated rather than areas where accidents have already been reported thanks to the near-miss data, which generates a map of hazards that did not previously exist.
Observing how these deployments are going gives the impression that Canadian cities are approaching AI traffic management with a practical caution that sets it apart from more ambitious smart city projects that promised complete transformation and presented challenging data governance issues. Intersections, signal timing, and safety assessments are the specific areas of attention. Instead than replacing the current infrastructure, the technology is being added to it. The collaborations, such as the Google Green Light partnership with Quebec City, are focused on cost-effective, scalable enhancements rather than complete infrastructure replacement. Its power may lie in its modesty. The same inclination toward expanding upon what already exists rather than beginning from scratch is evident in Innovation UBC’s research into camera and radar-based traffic monitoring.
Cities like Kirkland and Quebec City are producing the cleaner before-and-after data, but Vancouver, Toronto, and Montreal have their own initiatives. It’s difficult to ignore the fact that the cities most obviously promoting these experiments are not the biggest ones. Cleaner data, fewer conflicting variables, and more readable results are produced by smaller deployment contexts. It is actually unclear whether the results apply to the intricate interchange zones surrounding Montreal or the extensive artery network of downtown Toronto. In the hallways where it has been tested, the technology functions. Whether it operates at the scale that urban gridlock truly exists is an issue that will be answered over the next years.
