The UK Deployed AI Cameras That Watch Inside Every Car, and One in Five Drivers Called It an Invasion
What happened
In April 2024, ten police forces across the United Kingdom began testing an AI camera system mounted inside vans, designed to detect drivers using mobile phones and passengers travelling without seatbelts. The system, developed by Acusensus and operated by AECOM and National Highways, uses computer vision and object recognition to flag violations in real time as vehicles pass. Images and data captured by the vans are transmitted to police officers, who then decide whether to issue a penalty.
The rollout drew immediate pushback. A poll conducted by Confused.com found that 21 percent of motorists considered the system an invasion of privacy, even while acknowledging it would likely make roads safer. That pairing, accepting the safety case while rejecting the surveillance method, captures the core tension the deployment surfaced. Roads are public, but the interior of a car has long been treated as a private space, and these cameras watch both.
The enforcement stakes are not trivial. Drivers caught using a phone behind the wheel or travelling without a seatbelt face fines of up to 2,500 pounds. That level of consequence means the system is not merely observational. A camera with the authority to initiate a 2,500-pound penalty is operating with real legal weight, and how it identifies, records, and transmits evidence matters far more than a similar system running at lower stakes.
The objections were not about safer roads but about what the system observes and where that information goes. When a camera in a van captures images of a driver's hands, face, and vehicle interior, and transmits those images to an officer making a judgment call, it is doing something qualitatively different from a fixed speed camera reading a license plate. The data is richer, the identifiable detail is more personal, and the decision to act rests entirely with an individual reviewer working from footage taken without warning or consent.
The gap the backlash points to is not just about surveillance as a feeling but about what can be verified, after the fact, regarding how any given image was handled. A provable record of what a system did with the data it collected, how long it retained images of people who were not penalised, and what oversight governed that retention would give regulators and the public something concrete to examine. Without that record, assurances that the system operates responsibly rest on institutional trust rather than on anything anyone can actually inspect.
Reported impact
- Affected parties
- Not publicly disclosed
- Harm type
- Not publicly disclosed
- Scale
- Not publicly disclosed
- Financial impact
- Not publicly disclosed
- Regulatory action
- Not publicly disclosed
Classification
Relevant governance controls
Governance control mapping is not available for this record.
- No controls mapped
Not publicly disclosed
Control mapping is analytical. It does not state that any control would have prevented the incident.
Sources and evidence
This record was researched and written by the Index. The event is also catalogued in the following database, which is listed for cross-reference.