Speed ​​camera app developers face abuse from UK drivers | Road safety

The developers of a new app that uses AI to estimate the speed of a passing car say they were forced into anonymity by the vicious response from drivers.

The app, Speedcam Anywhere, is the product of a team of artificial intelligence scientists from Silicon Valley companies and top UK universities. Its creators hope it will inspire police to take speeding more seriously and allow residents, pedestrians and cyclists to document traffic offenses in their area.

But since its launch in March, the vitriol imposed on the team is such that they are afraid to share their true identities. “We get some pretty abusive emails,” said Sam, the app’s founder, who spoke on condition of anonymity. “It’s a Marmite product – some people think it’s a good idea, others think it turns us into a surveillance state.

“I can see both sides of this, but I think if you’re going to have speed limits, then it’s the law that you obey them, and you have to enforce the law. It’s not a personal vendetta against anyone, it’s just – how do we make our roads safe? There are 20,000 serious injuries on the roads each year, how can we reduce them? And the way we reduce them is to deter speeding.

An online review of the app said: “In East Germany, citizens were encouraged to report their neighbors to the Stasi for even the smallest societal offence. “Congratulations” on creating a modern version of this. If you couldn’t tell, I’m being sarcastic. This app disgusts me.

The application encountered difficulties. Google refused to allow the team to post it on the Play Store, saying it was not possible to estimate the speed of a passing vehicle using only AI – a claim turned out to be false when the company provided a demonstration of the technology. An iOS version has also been developed but Apple has not yet approved its distribution and given no reason for the delay. “We don’t know why they would block useful technology, something that could save people’s lives,” Sam said.

For several years, police forces have been accepting user-uploaded images of traffic crimes. This has allowed some citizens, like cyclist Mike van Erp, to submit evidence leading to hundreds of dangerous driving lawsuits. “What we’ve done is extend the kind of capabilities of dashcam systems, so you can automate the forensic video analysis that dashcams already do. So instead of a human watching a video for determine the violation, we have created software that automates the process.”

The app cannot cause drivers to receive speeding tickets. Since Speedcam Anywhere’s algorithm has not been approved by the Home Office, it is not legally a speed camera and cannot provide sufficient evidence for a police force to engage prosecution for speeding, although the broader offense of “dangerous driving” may apply if the driving is sufficiently careless.

Sam hopes the widespread use of the app will instead alert police to speeding hotspots and encourage them to take more action to prevent dangerous driving. “I think it’s a step in the big journey of how we make our roads safer and more accessible for everyone.

“Having roads too dangerous for children to cycle to school, roads too dangerous for parents to let their children cross – I think that’s wrong, and society has to get over it. Give back safer roads, make them less unpleasant, and then we can start looking at how else we can get around.

Apple has been contacted for comment.

How Speedcam Anywhere works

  • A user of the app opens it when he hears a car approaching at high speed and films the passing car.

  • The app uses the number plate of the passing car to search the DVLA’s public registration database to find the make and model of the car.

  • From there, it determines the distance between the axles of the car and compares it with the images to calculate the speed.

  • The user then has the option of recording the video or generating a report from it to share with the authorities.


Leave a Reply

Your email address will not be published. Required fields are marked *

Back to top