Road Safety Could Be Improved With AI-Powered Traffic Cameras: Study

Deploying traffic cameras powered by machine learning could prevent nearly $1 million in property losses from accidents each year.

A new study suggests that AI-powered traffic cameras could promote safer driving behaviors and significantly reduce the number of auto accidents on U.S. roadways, when compared to areas that either do not have cameras, or only have conventional camera systems.

In a report published in Management Science on April 15, researchers indicate that AI-equipped cameras with continuous video recording can dramatically improve road safety without causing unintended negative consequences.

Despite constant efforts to improve vehicle safety and roadway designs, automobile accident fatalities in the U.S. have remained alarmingly high in recent years, with more than 40,000 people killed annually since 2021, according to data from the National Highway Traffic Safety Administration (NHTSA). Following years of declining numbers, roadway deaths have surged since the pandemic, largely driven by increases in speeding, impaired driving, and distracted behavior.

In response, many cities have increasingly turned to automated traffic enforcement over the last decade. However, the impact and effectiveness of traffic cameras continue to spark debate among policymakers, researchers and the public.

In the new study, an international team of researchers examined the impact of traffic cameras on road safety using longitudinal data from a metropolitan city in China. 

Led by Zhi (Aaron) Cheng, the research team deployed advanced AI-powered cameras and conventional cameras at road intersections in a staggered rollout. The AI cameras used machine learning to detect traffic violations and record continuous video, while traditional systems relied on triggered image captures and are limited to specific types of infractions.

The study revealed that intersections equipped with advanced AI-powered traffic cameras experienced a significant and lasting drop in accident rates, outperforming areas with no cameras or only conventional systems. Notably, this reduction did not come at the expense of surrounding intersections, meaning crashes were not simply displaced elsewhere. 

Based on their analysis, the researchers estimated that deploying AI-enabled cameras across an entire city could prevent approximately 1,190 accidents and 496 injuries or fatalities each year, while also avoiding nearly $1 million in annual property damage.

The team identified three primary reasons for the effectiveness of AI-powered enforcement. First, the cameras offer superior technical capabilities, using machine learning to detect a wider range of traffic violations than conventional systems. Second, their continuous, real-time video recording enables better identification of crash causes and reinforces accountability. Third, the presence of these cameras serves as a behavioral deterrent, making drivers more aware of their actions and encouraging safer habits over time. 

According to lead author Zhi (Aaron) Cheng, these findings highlight the potential for AI-powered traffic cameras to serve as a practical and scalable tool for improving road safety and saving lives.

Rise in Pedestrian Fatalities

Traffic fatalities in the U.S. have become a growing issue of concern for both the government and researchers. Pedestrian fatalities increased by nearly 50 percent between 2014 and 2024, which is seven times faster than the growth rate of the U.S. population.

Experts have theorized that one possible factor for rising fatalities rates are the increased height of vehicle hoods, which significantly raises the severity of pedestrian injuries in accidents. Between 2010 and 2023, SUV presence on U.S. roads grew by 57% and pickup trucks became 13% heavier, both vehicle types with notably high hoods.

The rise in fatalities prompted the Insurance Institute for Highway Safety (IIHS) to propose a “30×30” plan, aiming to reduce traffic deaths by 30% by 2030 through lower speed limits, increased seat belt use, and reduced legal blood alcohol limits for drivers.