Across the United States, a new generation of traffic cameras is learning to read what drivers are doing with their hands, not just how fast they are going. Instead of relying on an officer’s quick glance through a windshield, artificial intelligence is now scanning high resolution images for glowing screens, loose seat belts and other telltale signs of distraction. The result is a quiet but sweeping shift in how road rules are enforced, with automated systems poised to catch phone use that used to slip by unnoticed.
These systems are already operating or being rolled out in multiple states, and they are built to do more than snap a license plate. They pair specialized camera rigs with machine learning models that flag likely violations for human review, turning every monitored lane into a kind of always-on checkpoint. Supporters frame the technology as a necessary response to stubborn crash numbers, while critics see the early architecture of a nationwide surveillance grid.
How AI traffic cameras actually catch your hands
The core idea is simple: instead of hoping an officer spots a driver hunched over TikTok in a moving car, roadside units capture a stream of images and let software do the first pass. In deployments described in recent reporting, the AI analyzes each frame to see whether a driver appears to be looking at a phone or not wearing a seat belt, then assigns a confidence score before any citation is considered. One system used in several states is designed to work from overhead gantries or trailers, peering through windshields at highway speeds and even in low light, so a quick glance at a messaging app on a modern SUV’s infotainment screen is no longer invisible to enforcement.
Vendors pitch this as a way to scale up enforcement without flooding roads with patrol cars, and the workflow reflects that logic. The software filters thousands of images down to a smaller batch that likely shows a violation, which is then reviewed by trained staff or police before a ticket is mailed to the vehicle owner. In one widely cited description, the AI is explicitly set up to flag only those photos that probably show drivers holding a device, leaving the final call to humans so that officers are not dispatched or citations issued without clear visual confirmation. That hybrid model is meant to blunt concerns about false positives while still letting the cameras quietly watch every passing pair of hands.
From Arkansas to Minnesota, a multi‑state rollout
What started as a pilot is now spreading across the map. In Arkansas, the Arkansas Department of Transportation has embraced an AI system that can monitor drivers for phone use and other violations as they pass through monitored zones. Officials there have described how the technology alerts drivers to its presence before they enter these areas, signaling that distracted driving is no longer a low risk gamble. The same state features prominently in broader coverage of America’s new AI traffic cameras, which explains how the system can watch a driver’s hands and catch them using a phone even when an officer is nowhere nearby.
Further north, Minnesota has become an early adopter of AI enforcement on busy corridors. On Highway 7 in the west metro, local leaders have turned to an automated camera setup as a new tool to make driving safer, with video showing how the system scans passing vehicles for drivers using cell phones. Separate reporting on AI cameras in five states notes that in Minnesota, one of the deployments is also being used to spot vehicles parked illegally in bus lanes, underscoring how the same hardware can be repurposed for multiple traffic problems once it is in place.
The Acusensus model and Connecticut’s big bet
Behind several of these deployments is Acusensus, an Australian born company that has turned its distracted driving platform into a major American export. The firm’s technology, marketed for both phone detection and speed enforcement, is at the center of a landmark contract with Connecticut. According to corporate disclosures, Acusensus Limited, listed on the ASX under the ticker ACE, has secured its largest United States deal to date, a work zone speed enforcement program worth $22.6 m. The same agreement is described as a US$22.6 million contract that will serve as a key foundation for the company’s US operations, signaling that state agencies are willing to invest heavily in automated enforcement infrastructure.
Connecticut’s interest is not limited to a single corridor. State officials have framed the work zone program as part of a broader push to slow drivers where crews are most vulnerable, and the Acusensus platform is designed to integrate with existing roadside equipment. A separate summary of the same deal notes that Acusensus wins US$22.6M for automated speed enforcement in Connecticut, reinforcing the scale of the commitment. For drivers, that means work zones in the state are increasingly likely to be guarded by AI assisted cameras that can clock speed and, in other contexts, scrutinize what is happening in the cabin.
California’s layered experiment with automation
On the West Coast, California is turning into a laboratory for automated enforcement at scale. Under new laws, the state has approved mailed civil tickets using red light and speed cameras, with a rollout that began at the start of this year. An official explainer notes that starting January 1, 2026, California will begin issuing these citations under a framework that treats them as civil penalties rather than criminal offenses, a distinction that shapes how drivers can contest them and how revenue is allocated. The same material is being promoted under campaigns like DriveSmartCA and RoadSafety, signaling that the state wants residents to see the cameras as part of a broader safety push rather than a pure cash grab.
Within California, Oakland is running a Speed Safety Cameras Pilot Program that shows how cities are tailoring the technology to local streets. The project overview describes speed safety cameras as a proven, life saving tool to reduce injuries and fatalities, and lays out specific corridors where devices will be installed, including stretches from Blake Dr to Gould St. Oakland’s plan emphasizes equity and transparency, with public maps and clear signage, but it also normalizes the idea that neighborhood arterials can be continuously monitored by automated systems. As California’s broader camera laws take effect, these city pilots are likely to serve as templates for other urban areas.
Arkansas tightens the rules, and police stay in the loop
Back in the South, Arkansas has paired its AI rollout with a political push to crack down on phone use behind the wheel. Coverage of the state’s new rules explains that Arkansas has tightened control over cell phone use in moving vehicles, and that the enforcement muscle increasingly comes from Acusensus hardware. Returning to the States, Dave Parker from ARDOT told Carscoops that Acusensus relies on AI to spot photos that probably show violations, then routes those images into a process that determines how the data gets used. That description underscores a key selling point for the technology: it promises more consistent enforcement without fully removing human judgment.
Another detailed account of the same system notes that back in the US, Dave Parker of ARDOT explained how the platform works in practice. According to his description, Acusensus uses AI to identify which images are worth a closer look, but it does not automatically call police without clear visual confirmation. Instead, flagged photos are reviewed before officers are alerted, and the rules in the U.S. are structured so that the AI cameras alert nearby officers, who then decide whether to pull a driver over. That approach, described in a broader analysis of how AI is watching you drive, is meant to keep a human in the loop even as the cameras quietly scan every passing car.
More from Morning Overview