London's facial recognition trial scanned 500,000 faces and caught no one
500k facial scans at UK stations yield no arrests, 1 false positive
A six-month live facial recognition trial across London railway stations scanned more than half a million faces, cost taxpayers £320,786 and nearly 100 police hours, and produced exactly one alert — a false positive that led to no arrest. British Transport Police is now extending the trial to London Underground stations, even as critics call the technology a serious invasion of privacy.
Success in a shop means no such people coming in. Success for the police trying to catch people means people being caught. In that respect the trial doesn't appear to have been very fruitful.
- balderdash
my mom lives in small town that only has two roads in/out of town and the police have alprs on those. there's no violent crime there, and all the police department does is traffic (DWI's, speeding and parking tickets, and accidents) and respond to small town crime (noise complaints, stolen bicycle, stray/lost dogs etc). I think the only thing that the alpr has ever caught people for is out of date vehicle registrations - but now the police department wants a flock camera because they are "great crime fighting tools"
- blitzar
16 deployments for 6 months at rail stations - they must have been off 99% of the time to only scan 500k faces.
For scale 500k people is 2 days of Liverpool Street Stations daily average
- jwally
I was thinking about this with flock cameras the other day...what's our actual ROI? Was the world horrible before and these initiatives are going to close a monster crime gap or is this just a giant fishing expedition ... just because?
- benbristow
I've had an idea in the back of my head but I'm no businessman so have this one for free.
Build an AI model, train it from the catalogues of JD Sports and Sports Direct (and other brands like Stone Island etc.) and how to detect balaclavas, illegal electric bikes and e-scooters. Have it link to cameras and do a percentage chance using machine learning on chance of crime. Also increase the percentage if a 'wheely' is done (riding only on back wheel).
Would minority report the majority of crime in this country in one fell swoop.
Could call it ScroteScanner.
Some would call it discrimination but the majority of ASB I see around my area are people dressed/using those exact things.
If they start to catch-on (unlikely) then worse case we have a fancier looking society with gangsters in suits again.
- whatever1
Man we technologists are so bad at dealing with societal issues.
We try to reinvent something that claims that keeps tabs on everyone and for sure will punish you if you don’t behave well but will reward you if do!
The thing has a name. It’s the religion.
- ungreased0675
I’m very against this type of pervasive surveillance, but that number seems implausible. The police may have only loaded a small group of suspects into the system, only targeted specific individuals, or some other reason we’re not aware of. The police protecting some criminals wouldn’t be a departure from standard practice in the UK.
- Frieren
The cameras in UK are there to remember citizens that they are watched. They are there for when there are peaceful protests to remember citizens that they are watched. Surveillance state created by cameras on the streets and tracking on everybody's phones is perfectly ok for our rulers.
I hear again and again that China is "bad" but I see that western countries copy everything from them... except going after big corporations. We will gather personal data and have files for each citizen, punish peaceful protests, use social media to control the masses but punishing the rich like China does is unacceptable. That seems to be the red line that the West will not cross, the one that protects billionaries.
- Someone
I don’t think it’s fair to conclude this doesn’t work at all.
The reason they did not get any arrests could be that sought-for criminals now avoid those stations. Given that there are relatively few of such criminals, that’s hard to prove or disprove (traffic would go down, but only by a slight amount)
It also could mean such criminals already avoided those stations, to start with. Again, hard to prove.
It also might mean people misbehave less now at these stations. So, did the number of incidents at such stations go down?
That, I think should be easier to (dis)prove, but also doesn’t require running live face recognition. Recording and only looking at recordings after an issue is reported would be the better option.