‘La tecnologia del riconoscimento facciale mi ha scambiato per uomo desiderato’
https://www.bbc.co.uk/news/articles/cqxg8v74d8jo
di F0urLeafCl0ver
‘La tecnologia del riconoscimento facciale mi ha scambiato per uomo desiderato’
https://www.bbc.co.uk/news/articles/cqxg8v74d8jo
di F0urLeafCl0ver
7 commenti
We complain the police don’t do enough, then the police do something and we say it is overkill…
I hate the public so much sometimes.
Have the police, without facial recognition tech, ever detained someone who wasn’t their target/suspect?
“I want structural change. This is not the way forward. This is like living in Minority Report,””
Except this is nothing like the minority report, the individual who they suspected this man to be was wanted for an offence already committed, and then once confirmed he was not the individual was released.
So I assume the structural change is what? Not attempting to correctly identify anyone wanted in case they are not the person wanted… Clearly facial recognition is not 100% accurate but here’s the kicker officers ability to recognise people walking the street is not 100% accurate either but I’d put money on the facial recognition technology being better.
A bit of A and B here.
It’s bold and described as orwellian in application. Is the tech good enough not to generate significant false IDs?
But
A copper staring at a street full of people can still misidentify someone. Usually this is cleared up by the person showing they are not the person that is wanted. At this point if the person refuses to confirm who they are an arrest is possible. Simple answer is if you get asked who you are, say it and move on with your life. No copper is going waste more time with you than they need to.
Facial recognition technology is not evidential in itself—it provides a score indicating the likelihood of a match, rather than a definitive identification.
True facial recognition—where accuracy can be reasonably assured—typically occurs within a metre of the subject. What we see with CCTV systems is far less precise: a face may merely pass a low confidence threshold before being flagged for operator review. If that operator lacks experience in identifying individuals from ethnic groups other than their own, the chances of error increase significantly.
For example, ask the average white person to describe a non-white individual, and the description is often vague or lacking in distinguishing detail. This highlights the need for proper training and assessment of personnel using facial recognition systems. Without this, the risk of misidentification and bias remains unacceptably high.
If only he’d had a digital id card, could have all been sorted in no time at all.
He could have just showed ID or provided fingerprints earlier I guess, rather than wait and argue for 30 minutes?