Drivers on major roads in Sussex are being watched by artificial intelligence. A camera system capable of detecting handheld phone use and seatbelt violations has been operating on the county’s motorways and A-roads since February 2024, part of a broader enforcement trial led by National Highways and multiple police forces across England.
The trial launched on 19 February 2024 and was scheduled to run through March 2025. Sussex Police was among the forces that participated. With the scheduled trial period now behind us, the question facing transport authorities is whether the technology proved reliable enough to justify a permanent, nationwide rollout.
How the system works
Cameras mounted on overhead gantries and roadside infrastructure capture high-resolution images of passing vehicles. AI software analyses each image, looking for telltale signs: a driver’s hand raised to their ear, a phone held at the wheel, or the absence of a seatbelt across the chest.
When the system flags a potential offence, the image is not used to issue an automatic penalty. Instead, it is routed to a trained human reviewer who examines the evidence and decides whether enforcement action is warranted. That two-stage process, AI detection followed by human verification, was built into the scheme from the start to reduce false positives and ensure the standard of evidence could hold up in court.
The approach allows enforcement at a scale that would be impossible with officers stationed on bridges or in patrol cars. A single camera installation can monitor thousands of vehicles per hour, flagging only those where the AI identifies a likely violation.
The trial built on earlier pilot work. A similar scheme was tested in the Coventry and Warwickshire area beginning in 2022, which explored AI-assisted detection for distracted driving and helped refine the technology before the wider rollout.
What the penalties look like
The cameras do not change the law. They change the odds of getting caught. Using a handheld mobile phone while driving in England carries a fixed penalty of six points on the driver’s licence and a £200 fine. If the case goes to court, the fine can rise to £1,000. For drivers who passed their test within the previous two years, six points means an automatic revocation of their licence under the New Drivers Act 1995.
Failure to wear a seatbelt carries a fine of up to £500, though fixed penalties of £100 are more common. Both offences are well-established in law, but enforcement has historically depended on a police officer physically observing the violation, a method that catches only a fraction of offenders.
Results and what remains unknown
As of May 2026, National Highways has not published a comprehensive evaluation of the trial’s outcomes through the sources reviewed for this article. No official detection counts, conviction rates, or formal accuracy assessments for the AI system have been identified in publicly available reporting. It is possible that results or audits have been released through channels not captured in this review, but none have surfaced in the material examined. That gap is significant. Without outcome data, it is impossible to say with confidence how many drivers were caught, how often the AI flagged false positives, or whether the cameras had a measurable deterrent effect on phone use and seatbelt compliance.
Sussex Police has also not disclosed detailed information about where cameras were positioned, how many were active at any given time, or what proportion of the county’s road network fell under surveillance during the trial. The deterrent value of the system depends partly on visibility: drivers who know cameras are present may change their behaviour, while those unaware of the monitoring may not.
The AI’s reliability under real-world conditions remains an open question. Rain, low light, heavy traffic, reflections on windscreens, and passengers holding phones near the driver could all create ambiguity in captured images. The human review layer provides a safeguard, but if the AI generates a high volume of false alerts, the system becomes less efficient and more labour-intensive to operate.
Privacy and the scope question
The cameras photograph vehicle interiors, which puts them in different territory from standard speed cameras or number plate recognition systems. That distinction has drawn attention from civil liberties groups. Big Brother Watch, a UK-based privacy campaign organisation, has previously raised concerns about the expansion of roadside surveillance technologies, arguing that the normalisation of AI monitoring on public roads risks eroding expectations of privacy.
Key questions remain about data handling: how long images are stored, who has access to them, and whether footage of drivers not committing an offence is deleted promptly. No detailed data retention policy specific to this trial has been published in the sources reviewed.
There is also a broader concern about mission creep. The cameras in their current configuration target two specific offences. But the underlying technology, AI capable of analysing the interior of a moving vehicle, could theoretically be adapted for other purposes: detecting smoking, eating, or even identifying occupants through facial recognition. Whether that ever happens depends on policy decisions and regulatory frameworks that are still taking shape, but the infrastructure now exists.
Where this fits in a global picture
The UK is not the first country to deploy AI cameras against distracted driving. New South Wales in Australia launched mobile phone detection cameras in 2019 and expanded them statewide by 2020. Early results there were striking: in the first year, the cameras caught more than 100,000 drivers using phones. The Australian model operates on a similar principle, AI flags potential offences and human reviewers confirm them, and has been credited with contributing to a measurable reduction in phone-related incidents on monitored roads.
That precedent matters for the UK debate. If the Sussex trial and its counterparts across England produced comparable detection volumes, the case for permanent deployment becomes harder to dismiss. If the numbers were modest, critics may argue the investment is better spent on conventional policing or road design improvements.
What comes next for English roads
The trial’s conclusion leaves National Highways and the Department for Transport with a decision. The data collected across Sussex and other participating areas during those 13 months will shape whether AI-assisted enforcement becomes a permanent fixture on English motorways or is shelved as a promising but impractical experiment.
Government road safety statistics underscore the stakes. According to Department for Transport data, in 2022 a driver using a mobile phone was a contributory factor in 20 fatal collisions and 91 serious injury collisions on British roads. Those figures almost certainly undercount the problem, since phone use is difficult to prove after a crash unless the device is examined.
For drivers in Sussex and beyond, the practical reality has already shifted. Even without a formal announcement of permanent deployment, the trial demonstrated that the technology works at an operational level: cameras can be installed, AI can flag offences, and human reviewers can process the results. The infrastructure and the institutional knowledge now exist. Rolling it back would be a deliberate choice, not a technical limitation.
Anyone who has grown accustomed to checking a notification at 70 mph, confident that no officer is watching, now has reason to reconsider. The officer may not be watching. But the camera almost certainly is.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
