5 ms·

"...is known for coauthoring a groundbreaking paper that showed facial recognition to be less accurate at identifying women and people of color, which means its

by 0-_-0 6y ago

"...is known for coauthoring a groundbreaking paper that showed facial recognition to be less accurate at identifying women and people of color, which means its use can end up discriminating against them" Wait, isn't that the other way around? If it can't recognize people of some category then it can't be used to discriminate against them, e.g. the police can't use it to identify peaceful protesters with those characteristics. I wonder what would have happened if the these networks were much better at recognizing women and people of color, would the paper then be about Google designing technology to detect minorities?

threatofrain 6y ago
When police error in identification can mean getting shot, black people deserve to be identified as accurately as individuals of any other race.
- Dylan16807 6y ago
  That depends enormously on the type of error, so you can't say that as a blanket statement at all.
frobozz 6y ago
If face detection defines them as not a person, then things that rely on there being a person in the field of vision will not work for them. (Like the racist soap dispenser that went viral a few years ago) If face recognition makes the old racist "they all look the same to me" declaration, then peaceful protestors get arrested for looking like criminals.
- 0-_-0 6y ago
  That's a straw man though, no classification system assigns a "criminal" score. Here's another quote that points out exactly what I was talking about: “There are two ways that this technology can hurt people,” says Raji who worked with Buolamwini and Gebru on Gender Shades. “One way is by not working: by virtue of having higher error rates for people of color, it puts them at greater risk. The second situation is when it does work—where you have the perfect facial recognition system, but it’s easily weaponized against communities to harass them. It’s a separate and connected conversation.” [0] [0] https://www.technologyreview.com/2020/06/12/1003482/amazon-stopped-selling-police-face-recognition-fight/ https://www.technologyreview.com/2020/06/12/1003482/amazon-s...
  - frobozz 6y ago
    I don't mean physiognomy, though such things periodically arise and are roundly and rightly ridiculed. I mean "Our computer says we have footage of you robbing a supermarket" (higher error rates) Another way for issues to arise (your easily weaponised point, above) is if you need a separate system for recognising people from certain groups. If it's all the same system, it's harder to argue that you are acting in good faith when you only follow up on matches on marginalised groups.
- wooger 6y ago
  "Peaceful protestors" became a bit of a joke this year, as newsreaders stood in front of burning cars & buildings telling us a peaceful protest was going on.
htfu 6y ago
> e.g. the police can't use it to identify peaceful protesters with those characteristics It can be used to (mis) identify them if the higher error rates are accepted. The problem isn't really the tech, but its application.

by 0-_-0 6y ago

threatofrain 6y ago

When police error in identification can mean getting shot, black people deserve to be identified as accurately as individuals of any other race.

Dylan16807 6y ago
That depends enormously on the type of error, so you can't say that as a blanket statement at all.

frobozz 6y ago

If face detection defines them as not a person, then things that rely on there being a person in the field of vision will not work for them. (Like the racist soap dispenser that went viral a few years ago) If face recognition makes the old racist "they all look the same to me" declaration, then peaceful protestors get arrested for looking like criminals.

0-_-0 6y ago
That's a straw man though, no classification system assigns a "criminal" score. Here's another quote that points out exactly what I was talking about: “There are two ways that this technology can hurt people,” says Raji who worked with Buolamwini and Gebru on Gender Shades. “One way is by not working: by virtue of having higher error rates for people of color, it puts them at greater risk. The second situation is when it does work—where you have the perfect facial recognition system, but it’s easily weaponized against communities to harass them. It’s a separate and connected conversation.” [0] [0] https://www.technologyreview.com/2020/06/12/1003482/amazon-stopped-selling-police-face-recognition-fight/ https://www.technologyreview.com/2020/06/12/1003482/amazon-s...
- frobozz 6y ago
  I don't mean physiognomy, though such things periodically arise and are roundly and rightly ridiculed. I mean "Our computer says we have footage of you robbing a supermarket" (higher error rates) Another way for issues to arise (your easily weaponised point, above) is if you need a separate system for recognising people from certain groups. If it's all the same system, it's harder to argue that you are acting in good faith when you only follow up on matches on marginalised groups.
wooger 6y ago
"Peaceful protestors" became a bit of a joke this year, as newsreaders stood in front of burning cars & buildings telling us a peaceful protest was going on.

htfu 6y ago

> e.g. the police can't use it to identify peaceful protesters with those characteristics It can be used to (mis) identify them if the higher error rates are accepted. The problem isn't really the tech, but its application.