New York City to Install 80 More AI Sensors Tracking All Street Users at 100 Intersections

The New York City Department of Transportation will expand its street activity sensor program to about 100 locations across the five boroughs. The department first tested the sensors at 20 intersections during a pilot that began in 2023 and now plans to add about 80 more sites. The sensors use computer vision AI to classify pedestrians, cyclists, cars, trucks, buses or scooters.

They are mounted on city street infrastructure such as poles or signs. The devices can measure speeds, capture turning movements and map how people move through a street or intersection. Video frames are deleted nearly instantly after the sensor collects the count.

NYC DOT says only anonymous data is kept and that faces and license plates are deliberately obscured. Some information from the sensors will be added to New York City's open data page. Traditional traffic studies often rely on workers counting vehicles for a few hours during limited periods.

The sensors collect data continuously, allowing officials to observe patterns across different times, weather conditions and seasons. The data can show where pedestrians cross mid-block because a crosswalk is too far away, where cyclists swerve around loading trucks, or where vehicles turn too fast near schools.

City planners can use those patterns to add crosswalks, adjust signal timing or redesign bike lanes before crashes occur.

Fox News reported that the sensors detect near-misses such as a car door opening near a cyclist or a delivery truck blocking a driver's view. Repeated close calls at the same location can prompt changes before injuries happen. The department states the system is designed to keep traffic patterns rather than personal identities.