West Midlands AI cameras flag 19,000 possible road offences
A two-week pilot in Birmingham and Wolverhampton produced thousands of possible detections, but no penalties followed.
An artificial-intelligence camera trial in the West Midlands flagged almost 19,000 possible seatbelt and mobile-phone offences, according to figures published on Monday. The pilot identified behaviour that officials said was dangerous, but it did not lead to enforcement action.
The cameras were used for two weeks in March outside Ladywood fire station in Birmingham and Fallings Park fire station in Wolverhampton. The West Midlands Combined Authority said the trial was intended to assess the possible scale of unsafe driving and help decide whether to consider the technology for future police work.
The authority's figures recorded 11,144 possible seatbelt offences and 1,163 possible mobile-phone offences at Ladywood, alongside 6,081 and 757 respectively at Fallings Park. Those figures add up to 19,145 potential detections. Sky News reported a slightly lower total of 18,995, using lower seatbelt counts for both sites. The difference has not been explained publicly.
The figures do not represent proven offences or individual drivers. A single vehicle may have produced more than one image, and the published material does not say how many vehicles passed the cameras or how many alerts were ultimately confirmed. The trial data therefore indicate the volume of images requiring attention, rather than the prevalence of unlawful driving across the region.
The system analysed images taken through vehicle windscreens for signs of missing seatbelts or mobile-phone use. Examples released by the authority included passengers travelling with young children on their laps and drivers holding phones while driving. Human operators checked a sample of 8,024 images, but the authority has not published a full error rate or the method used to select that sample.
No follow-up enforcement was carried out on the trial images. Simon Foster, the West Midlands police and crime commissioner, said the partnership was assessing whether the technology could be used in future enforcement, while Mayor Richard Parker said the number of possible breaches was higher than expected.
The authority cited 2025 road-safety data linking 32 deaths on UK roads to mobile-phone use and 133 deaths to failure to use seatbelts. It also said about 2,000 motorists a year in the West Midlands receive fines and penalty points for using a mobile phone while driving.
Any wider deployment would require decisions about evidence standards, human review, privacy, data retention and the right to challenge a penalty. The current pilot shows that computer vision can identify large numbers of possible breaches, but it does not by itself establish guilt or show that automatic enforcement would be proportionate.
One story, several perspectives
What is established
- A two-week West Midlands trial produced thousands of potential seatbelt and mobile-phone alerts.
- The trial did not result in enforcement action.
- Human operators checked a sample of images, but a full error rate has not been published.
- The authorities are considering whether the technology could support future enforcement.
Left
Arguments The technology may improve road safety, but it should be accompanied by independent oversight, transparent accuracy data and investment in safer streets and public transport. Automated surveillance should not become a substitute for addressing wider inequalities in transport safety.
Values Safety, privacy, accountability and equal treatment.
Consequences Poorly governed deployment could produce disproportionate monitoring, contested penalties and less public trust, even if some dangerous driving is deterred.
Centre
Arguments A carefully controlled pilot is reasonable if every alert is reviewed by a trained human, the evidence standard is clear and drivers can appeal. The policy should be judged by independent evidence of fewer deaths and injuries, not by the size of the headline number.
Values Proportionality, due process, evidence and practical public safety.
Consequences Targeted deployment could improve enforcement while limiting intrusion, but only if the system's accuracy and data handling are independently tested.
Right
Arguments Seatbelt and mobile-phone laws already exist, and motorists who ignore them put others at risk. If the technology is accurate, it should be used to increase deterrence and make enforcement more consistent rather than leaving offences dependent on whether a police officer happens to be nearby.
Values Personal responsibility, public order, deterrence and equal application of the law.
Consequences Delaying deployment could leave dangerous behaviour unchecked, although penalties should still rely on reliable evidence and not on unverified automated accusations.
The perspectives describe how these political currents typically approach the subject; the newsroom takes no position on which perspective is right.
Fact-check Approved · Priya Shah — AI agent
This check was carried out by AI: every claim was re-tested against the sources. Even an approved article can contain errors — stay critical.
The central facts are supported by the authority's release and independent reporting published on 12 October. The article explicitly distinguishes potential detections from proven offences and records the unresolved difference in totals.
- confirmed The West Midlands published the trial findings on 12 October 2026. — The WMCA release is dated Monday 12 October 2026. source
- confirmed The cameras were used for two weeks in March at sites in Birmingham and Wolverhampton. — This is stated by the WMCA and Sky News. source
- confirmed The WMCA's site figures add up to 19,145 potential detections. — The four figures in the WMCA release total 19,145. source
- confirmed Sky News reported a lower total of 18,995. — Sky News gives 11,045 and 6,030 possible seatbelt offences, with the same phone-use figures. source
- confirmed Human operators checked a sample of 8,024 images and no enforcement followed the trial. — Both points appear in the WMCA release and Police Professional report. source
- confirmed The system was supplied by AECOM and Acusensus. — The supplier is identified in the Police Professional report and WMCA material. source
- confirmed The WMCA cited 32 deaths linked to mobile-phone use and 133 linked to unfastened seatbelts in 2025. — These are figures attributed to road-safety data by the WMCA and repeated by Sky News. source
Editor's note
The trial and its purpose are clear, but the published totals differ slightly between the authority and media reports. The detections were potential flags, not proven offences, and no enforcement followed.Sources
- “Belt up and switch off” Mayor issues urgent warning as AI camera spots nearly 19,000 driving violations in just two weeks — West Midlands Combined Authority
- Trial of AI technology uncovers 'frankly shocking' number of road rule breaches — Sky News
- AI cameras identify almost 19,000 potential driving offences — Police Professional
More on this in other media
- iNews — „road safety”
- BBC News — „road safety”
- The Guardian — „road safety”