Auto journalists being harassed by police is simply a result of cameras working ‘correctly’

The story of The Drive’s Joel Feder being detained The hit case of false flak camera for a “stolen” vehicle has spread around mainstream media It has been like wildfire since it was published. Ever since this story came to light, many people have been asking how things happened this way and how to prevent such an incident from happening again. Last week Feder posted a follow-up story after speaking to Flock’s chief communications officer and the news is not good. It turns out that the system is working “correctly”, according to Flock, which is completely reassuring. Correct?

The original stop occurred because a production studio in California lost the license plate during a photo shoot and reported it to the Los Angeles County Sheriff’s Department (not the LAPD, as originally reported). New Jersey manufacturer plate 34 03 DTM was erroneously recorded as 34 DTM. It was also listed as stolen, not simply lost. This ultimately led to police in Plymouth, Minnesota, tracking Fedor via surveillance cameras for a week before cornering him and his wife in a parking lot due to a clerical error. Fedor’s not-stolen Range Rover press car had the plate 34 10 DTM, and those two middle digits are much smaller than normal numbers, so the cameras weren’t even taking it into account.

Fedor was fortunately able to explain himself to the satisfaction of the authorities, which led to the discovery of this error. He and one of the officers then realized that this situation could happen again to anyone driving one of Jaguar Land Rover’s fleet of cars, which has 34##DTM in its license plate. It didn’t take long for his prediction to come true.

Last week, Pickup Truck + SUV Talk’s Tim Esterdahl got himself pulled over while driving a Range Rover in the manufacturer fleet with license plate 34 08 DTM. It was exactly the same situation, a flak camera hit an allegedly “stolen” license plate. Esterdahl had read Fedor’s story and fortunately found it quite easy to explain the situation, but he ran into another problem. The officer was unable to verify the VIN of the Range Rover. Since the vehicle still belongs to JLR and has never been sold to the public, its VIN does not appear in the police computers.

After an hour on the side of the road, the officer let Esterdahl go despite the VIN problem not being resolved. Esterdahl is very clear in his video that he has no problem with the officer himself, what he did, or how he acted. He was genuinely confused by the information he was being given (and not being given) by the computers he relied on to do their jobs. And once again, it was all because of a wrongly registered license plate on the other side of the country. For a system that can supposedly track vehicles or even individual people without license plates, this is a huge mistake.

Flock Systems has finally addressed this problem. Flock Chief Communications Officer Joshua Thomas said drive: :

“The way ML (machine learning) works is it reads exactly what it’s supposed to read. It was fed the characters that you said, 34 DTM, and it spits back (the result) with the characters, 34 DTM,” Thomas said. “It asked, can you find it? And he found it. It didn’t say if there’s more here, don’t do it. It just said, is it there? And the answer was yes.”

So technically, the system is working perfectly when it comes up with all these false matches. There appears to be no difference between full and partial matching. This is a big mistake, because sometimes the police has to take only partial action. If the Plymouth Police had known that 34 DTM was only a partial match, their reaction would probably not have been as intense as that experienced by Federer. While Thomas agrees that there is room for improvement in this area, he also places the responsibility on the people who use it:

“A human entered the system. And a human did not enter enough information into the system,” he said. “The system read. And the result was a problem. But the steps to get there were full of human errors.”

Flock claims its system is 99% accurate. This may be true, but it is also true that it is Scanning 20 billion vehicles per monthWhich means it’s still making 200 million mistakes per month. While the situation surrounding JLR’s fleet vehicle plates is unique, it’s not even a drop in the bucket when it comes to the millions of mistakes in the system, combined with mistakes by law enforcement (which are only human, mostly). Fedor said it best in his most recent update:

Technology is moving so fast without guardrails and it’s increasing human errors, humans are not catching errors at the front or middle or back end of the loop, at least not in this case, and that can get someone hurt or even killed.

If flock cameras detect a falsely marked license plate in a bad neighborhood with someone of color behind it, things can get really bad, really quickly.