
Le telecamere di riconoscimento facciale sono troppo distorte da usare a Notting Hill Carnival, dicono attivisti | Polizia Stradale
https://www.theguardian.com/uk-news/2025/aug/16/facial-recognition-cameras-too-racially-biased-to-use-at-notting-hill-carnival-say-campaigners
di InternetProviderings
23 commenti
‘The letter says that since the Met announced its Notting Hill plan, a high court challenge has been launched by the anti-knife campaigner Shaun Thompson. A Black British man, Thompson was wrongly identified as a criminal, held by police, and then faced demands from officers for his fingerprints.’
Software that is known to have issues with identifying black people shouldn’t be used at all. It leads to a disproportionate targeting of innocent black civilians.
What nonsense. We should have facial recognition everywhere and a national database of every citizen. This technology works and it works very well (source: have lived in China). These people are just angry that criminals will be arrested.
> held by police, and then faced demands from officers for his fingerprints
> …when he was surrounded by officers and held for half an hour. He has likened the discriminatory impacts of LFR to ‘stop and search on steroids’,”
This can’t be true. We’ve been assured by officers here on reddit that it’s just a case of merely showing your ID and moving on, absolutely zero inconvenience to you at all.
I find it very interesting that black and Asian faces are thought to be a problem in the west, where the majority are white yet in China they have the opposit issue.
It’s almost like the minority are the ones still teaching the ai how to identify them.
As long as the human officers are given the info and can check the person against the photo before any action is taken where the issue?
The obsession with making this about race just muddies the water. Do not fall for it. They just want us all to fight among ourselves.
The real issue is that facial recognition is being rolled out on citizens without consent, without explanation, and without accountability. That’s the dead-cat strategy in action….keep everyone arguing about bias while the surveillance quietly becomes normal.
It’s not about whether the AI is biased. It’s about governments normalising a surveillance state without ever asking permission.
No harm in using it to track down suspects. It should be inadmissible as evidence though. We know you were there because facial recognition puts you there is no better or worse than a witness having put you there. Might form part of a picture but assuming it’s right is dangerous.
It’s because the alarm system would overload if it could recognise all the crims there.
Chicago recently got ride of a shot spotter system that allowed the police to pinpoint gunfire locations. It was removed because too many minorities were being arrested. It was too effective.
So from a technological standpoint, this is an interesting problem that has been around for a while.
In the past, and probably still with some devices today, camera software has been better in regards to balancing light and colour levels for people with light skin tones then those with darker ones. The reason, was test data. The camera software had been tested on more people with lighter tones, so the software was better calibrated for taking pictures of them.
Makes me wonder if we have a similar situation here. The test data used when making this software simply had more white men in the sample pool then woman or people of colour, and so is more accurate in identifying them.
If they’re not used and continually tested with the new software updates how would you know if they’re racially biased or not?.
There has been and is still ongoing, considerable research and updates to get this sort of thing right. Seems to me the Notting Hill carnival would be a great place to test it.
Having read the article , I fail to see any reason why it shouldn’t be used. If anything this would be a great opportunity to improve the technology and therefore remove the ‘racial bias’ outlined in the article
Classic Guardian article. The disproportionally high frequency of violence & death at this one specific event shouldn’t be anyone’s concern, vague campaigner claims of racism are the real problem.
If it takes the race card to get rid of these things then so be it…
Facial recognition systems should not be allowed at all, what a terrible invasion of privacy and constant surveillance.
Facial recognition cameras too effective at spotting criminals to use at Notting Hill carnival, say campaigners
Far cry from HP webcams not being able to detect a black face…
Hasn’t the “bias” been worked out now? I thought that’s what I’d heard. Would love some data on this.
What do people fear about facial recognition? I don’t do anything wrong so carry on recording me!
The fact that facial recognition is being rolled out at all is a crime…..
It could acually cause more harm, putting technology that has historically gotten more false positives with black people in a area thar is going to be filed with majority black people celebrating black culture could be a bad idea
Even if you remove the racial element from it the fact is that this country used to berate the likes of China doing this type of shit but here we are doing it ourselves
When a majority of crimes are commited by a certain minority, how does an AI system is meant not to be biased?
Pulling the race card would just make things worse for everybody.