Controversial Police Facial Recognition Test Fails to Recognize Anyone in London (independent.co.uk) 41
Police officers have just begun testing facial recognition software in London. Slashdot reader Bruce66423 reports:
After all the concern about the [first] trial, it appears to have been a bust. "Police have admitted that no one was arrested during a trial of controversial facial recognition technology, which sparked privacy and human rights concerns," reports the Independent. On the other hand, this may lead us to get to get complacent about the threat that is out there.
Detective Superintendent Bernie Galopin, the force's lead for facial recognition technology, pointed out that "All alerts against the watchlist will be deleted after 30 days and faces in the database that did not generate an alert were deleted immediately." But an advocacy and policy officer from the National Council for Civil Liberties complains that pedestrians were never informed what was happening -- except for one man who was apparently stopped erroneously after a "false positive" match (which the officers failed to first confirm on their own).
"Opponents argue that the software currently being used by British police forces is 'staggeringly inaccurate' and has a chilling effect on society," reports the Independent, "while supporters see it as a powerful public protection tool with the ability to help track terrorists, wanted criminals and vulnerable people....
"The use of facial recognition is more prevalent in the U.S., where it was used to track down an alleged mass shooter following a massacre at a newspaper's office last week."
Detective Superintendent Bernie Galopin, the force's lead for facial recognition technology, pointed out that "All alerts against the watchlist will be deleted after 30 days and faces in the database that did not generate an alert were deleted immediately." But an advocacy and policy officer from the National Council for Civil Liberties complains that pedestrians were never informed what was happening -- except for one man who was apparently stopped erroneously after a "false positive" match (which the officers failed to first confirm on their own).
"Opponents argue that the software currently being used by British police forces is 'staggeringly inaccurate' and has a chilling effect on society," reports the Independent, "while supporters see it as a powerful public protection tool with the ability to help track terrorists, wanted criminals and vulnerable people....
"The use of facial recognition is more prevalent in the U.S., where it was used to track down an alleged mass shooter following a massacre at a newspaper's office last week."
Re: (Score:3)
ARM... for now.
The British public seems quite willing to put perceived safety above freedom, which is unfortunate because politicians are good at scaring people.
Re: (Score:3)
ARM... for now.
ARM is a bit old to be considered part of "these days" though you could argue that they're still making new designs and it's those to which you were referring. However, almost none of the big players use the standard design any longer and instead license the ISA and build their own custom cores. Even some of the Chinese phone companies are starting to get in on building their own SoCs and it probably won't be long before they start developing their own cores for those designs as well.
Considering how much
Re: (Score:2)
" Every failure is a reason for more EU."
Oh like the American military? I see!
Re: Obvious solution: (Score:2)
Yeah, I miss the days when America wasn't a vassal colony of a foreign power :(
Counter (Score:1)
Wouldn't something as simple as a round clown nose counter this sufficiently?
Re: (Score:2)
Facial recognition works on that trapezoidal area between the top of the eyebrows, nose and mouth. They look for the relative lengths between different points on each side of the face; corners of the eyes, eyebrows, mouth, nose. Facial hair and shaved eyebrows are enough to throw the system off. Same with sunglasses, long hair and a floppy hat.
Re: (Score:2)
AI facial recognitions works by extracting wealth from the public purse, via lying lobbyists and paid of politicians. How well does it need to work, not as well as the propaganda feed into corporate main stream media about how great it is. Why does it work so badly because profits first, the spend the absolute minimum amount possible, whilst charging the maximum amount possible, use propaganda channels to cover it all about because bribes are expensive and not tax deductible and it is all, absolutely 100% a
Good (Score:1)
Find more super-recognisers instead ? (Score:2)
Blasphemous as it may be to suggest this on a techie site ... maybe the money could have been better spent finding and training up more people who can recognise faces better than other mere mortals, seems a very freaky skill to someone like me who can barely remember anyone!
http://superrecognisers.com/ [superrecognisers.com]
Clickbait from the Editor (Score:1)
"The use of facial recognition is more prevalent in the U.S., where it was used to track down an alleged mass shooter following a massacre at a newspaper's office last week."
No, no it wasn't. The Maryland case was the police using facial recognition to identify a suspect they already had in custody, who refused to identify himself. Nothing whatsoever to do with tracking someone down.
Come on EditorDavid. If you're going to editorialize, at least try to get the facts straight. There might be some concern about the cops using this success to justify more surveillance, but in truth the one has nothing to do with the other. They just ran his picture through DMV photo databases, mugs
Too bad (Score:3)
That's a shame (Score:2)
I blame increasing dental visits. (Score:2)
I blame increasing dental visits.
The NHS just had an anniversary that was highly publicized, and it's clear people are finally doing something about their manibles, maxilla, and teeth, all of which would throw this sort of thing off.
Shocking (Score:2)
Nobody could have foreseen that a system with a 98% false positive rate in trials would fail to work properly when rolled out.
No arrests does not equal no one recognized (Score:2)
In the US and elsewhere, cell phone data is often used to exclude or include people who might be under suspicion of a crime. It doesn't necessarily lead to arrests, and may not be used in court, but it helps reduce the amount of work needed to be done during the process of an investigation. This data is usually used after the fact, not at the moment police are swooping in. Facial recognition likely falls into this category.
Just because the system didn't lead to any arrests, doesn't mean that no one was reco