AI traffic cameras are preventing thousands of crashes each year

A new line of research shows that artificial-intelligence traffic cameras can cut crashes far more effectively than earlier enforcement tools. These cameras curb collisions both where they are installed and in neighboring streets.

The findings come from a collaborative project led by scholars at the London School of Economics, Sun Yat-sen University, and the University of Wisconsin–Madison.

Making smarter intersections

Drawing on five years of detailed police and sensor records from a large Chinese metropolis, the investigators compared intersections upgraded with “advanced” AI cameras. They measured these against sites left with conventional triggered cameras or no monitoring at all.

The team modeled the spill-over effects across the broader road network to see whether accidents simply migrated to the next junction.

Advanced cameras differ from older devices in two crucial ways. First, they run machine-learning algorithms that can flag a wide spectrum of risky actions, from illegal turns to sudden lane changes, in real time.

Second, they stream continuous video rather than grabbing single frames only when a car breaks the light. That unbroken footage gives authorities the raw evidence needed to reconstruct crashes quickly and identify fault with greater confidence.

Cameras prevent damage and injury

Using a staggered event-study design, the team estimated that equipping every major intersection with the smarter cameras would have averted about 1,190 crashes a year.

The projection translates into 496 avoided injuries or deaths and roughly one million US dollars in property damage saved. These figures place the technology alongside far costlier road-engineering interventions.

Crucially, the researchers found no evidence that the reduction was offset by new accidents at nearby junctions. Earlier “red-light” systems often had this drawback, pushing reckless drivers into less monitored streets.

“Our study shows that AI-powered traffic cameras don’t just document violations – they promote safer driving behavior and significantly reduce accidents, even in nearby areas without cameras,” said lead author Zhi Cheng, a scientist at the London School of Economics.

“Traffic safety is a persistent and complex challenge. Our findings offer evidence that carefully deployed AI technologies can create real, system-wide improvements without unintended negative consequences,” added senior author Min-Seok Pang from the University of Wisconsin–Madison.

How AI changes driving

Behind the headline gains lie three overlapping mechanisms. The automated detection effect expands the range of punishable offenses, increasing the probability that dangerous habits will draw a fine.

Real-time recording boosts enforcement credibility by making it harder to contest citations or dispute liability.

Finally, the driver learning effect arises as motorists internalize the new surveillance environment; awareness of ubiquitous, accurate monitoring nudges them toward safer behavior well beyond the camera’s immediate field of view.

By isolating those channels, the study helps regulators grasp why earlier generations of fixed traffic cameras produced mixed or fading returns.

Systems limited to photographing vehicles that ran red lights could still miss speeding, tailgating, or illegal overtakes. They also captured only isolated frames, creating loopholes for legal challenges and leaving police with little context when crashes occurred.

Machine-learning cameras close those gaps – and do so at falling cost as computing power cheapens.

Smarter laws, safter streets

Worldwide road deaths have hovered at distressingly high levels despite advances in vehicle safety and public awareness campaigns. In the United States alone, annual fatalities exceeded 40,000 in recent years, erasing a decade of progress.

Cities hunting for scalable, budget-friendly countermeasures have begun to experiment with AI-enhanced monitoring, yet evidence has lagged behind enthusiasm.

The new study offers the strongest real-world evidence yet that advanced cameras improve safety without simply shifting risks elsewhere.

The authors caution that technology is only one piece of the puzzle. To maximize the benefit, municipal leaders need clear legal frameworks for data privacy, streamlined processes for adjudicating violations and outreach campaigns that explain how continuous video will be used and protected.

Well-designed warning signage and fine schedules can reinforce the deterrent effect while maintaining public trust.

AI traffic cameras are evolving

Although the dataset comes from a single Chinese city, the research design can be replicated elsewhere. Many jurisdictions already keep granular crash records and, increasingly, high-resolution traffic footage.

Integrating those archives with machine-learning analytics could reveal location-specific patterns, such as peak periods for illegal left turns. It could also identify corridors where distracted driving surges and guide the selective placement of traffic cameras.

The technology is also evolving. Newer platforms pair visual feeds with radar and lidar, allowing three-dimensional tracking of vehicles, cyclists, and pedestrians.

When combined with edge computing, these multisensor nodes could alert drivers in real time or trigger adaptive signal timing to diffuse congestion that often precedes collisions.

Safer roads by design

The promise of AI enforcement aligns with the “Vision Zero” movement that seeks to eliminate traffic deaths altogether. In that framework, every fatality is a system failure rather than an inevitable cost of mobility.

Advanced cameras, the authors argue, give planners a nuanced lever: they deter risky maneuvers, clarify accountability when crashes occur and gather continuous data to refine future interventions.

Alongside redesigning streets and improving public transit, ubiquitous machine vision may help cities create safer roads. The goal is to make serious injuries the exception, not the rule.

As artificial intelligence spreads from smartphones into asphalt and stoplights, it reframes the basic social contract of driving.

The study’s message is clear: when deftly deployed, algorithms can save lives not by replacing human judgment but by making reckless choices harder and careful habits easier.

The study is published in the journal Management Science.

—–

Like what you read? Subscribe to our newsletter for engaging articles, exclusive content, and the latest updates. 

Check us out on EarthSnap, a free app brought to you by Eric Ralls and Earth.com.

—–