AI pothole patrol to snap flaws in Britain's crumbling roads
Now if only the councils could afford to fill them
An oft-repeated myth is that potholes form through a combination of surface cracks, water, and traffic, but they're actually caused by chronic levels of underinvestment in public infrastructure.
Surrey County Council, the local authority for the region southwest of London, is bucking that trend.
Yes, Surrey "will now use AI to detect potholes to help speed up road repairs across the county."
"Computer vision cameras fitted to dashboards inside highways vehicles will spot and photograph potholes which will then be automatically recorded for repair. Future enhancements will see other defects such as missing signs and foliage overgrowth also programmed for repair," the authority said.
"Highways Inspectors will no longer need to step onto the road to manually carry out inspections which will make the process safer and more efficient. Potholes that don't need immediate attention will also be regularly tracked to ensure they are dealt with when needed."
Because using one's eyes and brain to take note of the location and state of the road is so 2010s. Joking aside, Surrey in particular stands to gain from some sort of solution. Research by automotive services company RAC found that of 18 councils which responded to a Freedom of Information request, Surrey County Council had the biggest increase in pothole claims, from 734 in 2022 to 3,418 in 2023.
What's worse, only 12 out of 1,204 claims for pothole damage were repaid by Surrey County Council last year, totaling £4,435.15. So residents' cars are sustaining heavy damage and many of them are angry enough to complain about it.
The UK does have a system for reporting potholes, where citizens are expected to pop their postcode into a website, which redirects them to the local authority responsible. We imagine, though, that the diffusion of responsibility kicks in here, with the net result being that motorists think someone else will already have reported the gaping chasm in the road, thus it doesn't actually get reported.
Or it does, it's just that the council doesn't have any money to fix them.
However, Surrey's partnership with Route Reports, the company supplying the technology, seems like it could make a difference, especially when it comes to spotting and cataloging the craters and other road faults.
- Apple solves broken news alerts by turning off the AI
- Google reports halving code migration time with AI help
- Nvidia shovels $500M into Israeli boffinry supercomputer
- UK businesses eye AI as the cheaper, non-whining alternative to actual staff
The county's road maintenance vehicles will be equipped with black boxes on the dashboard connected to a 4K, 60 fps camera on the front. Inside is an Nvidia chip designed for image processing, along with a GPS receiver and a 4G connection. This means that workers can keep driving when they come across a pothole. Instead of stopping, getting out, and inspecting the fault, the system has already done the most inconvenient part of the job.
It is able to "calculate the width, length, and depth of potholes it discovers – which means that those problems that are too small for a human inspector to bother with can be noted down before they become a problem, and maintenance teams have more data to help them prioritize," according to a deep dive by tech writer James O'Malley. And it can do all this even at speeds up to 60 miles per hour (96 kph), with accuracy of up to 95 percent compared to humans.
Route Reports CEO Connell McLaughlin said in a statement: "We're thrilled to formalize this long-term partnership with Surrey County Council, who were among our earliest supporters and collaborators in developing Route Reports' AI road maintenance platform.
"Their feedback and hands-on involvement played a key role in refining our defect detection technology to ensure it meets the real-world needs of the community. This partnership exemplifies how innovative solutions and local expertise can come together to make roads safer, more efficient, and more sustainable."
It's an example of real AI being genuinely and practically useful as opposed to the flood of hallucinating large language model chatbots being pushed at businesses in the hopes of replacing humans. ®