Amazon will temporarily stop providing its facial recognition software to police

“Amazon said Wednesday it will stop providing its facial recognition technology to police forces for one year, amid questions about the company's commitment to fighting systemic racism.“

I imagine that facial recognition software would aid the police in finding looters and the people that are beating and killing innocent bystanders but amazon appears not to care.

The reason amazon gives is as follows..

“ Amazon said it made the decision in hopes of giving Congress time to "implement appropriate rules" for police use of facial recognition algorithms, which the US government has found are often more likely to misidentify people of color.”


Correct me if I’m wrong but if their software can’t reliably identify people of color then it’s not very good software.
Just feels like another company pretending to take the moral high ground.

www.cnn.com...

a reply to: Bluntone22

I think now would be a good time to show displeasure with Amazon by not ordering from them. See it works both ways.

a reply to: Bluntone22
Boycott AMAZON!

Sounds like some good is coming out of all of this.

Let's hope it goes on for more than a year.

a reply to: Bluntone22

asking for a friend - Is Amazon saying POC's all look alike on the poorly light reflected images?

Wonder why that would be?

So pretty good of them to protect the non-pocs from being identified by the police.

a reply to: puzzled2

I don't know if it's matching facial features or if the dark skin messes with the machine.
I've heard that Asians confuse the software too so I would lean towards similar facial features.

Can Amazon glow any brighter?

Did they say where or how they have been using the software?

a reply to: Bluntone22

Looters tend to wear bandanas/scarfs that cover everything below their eyes. Really screws up facial recognition.

So, is Amazon saying that all black people look alike to it's software?

Sounds rather like a racist piece of junk to me (and I have been in the Software Engineering field for 25+ years). Why are they using outdated metric measurements? It's soooooo 2012. They may want to switch to using a convolutional layer at the start.

SMH

If the democats antifa and BLM crowd directed the looters and rioters to their amazon warehouses...See how fast they change their minds!

forestshipping.com...

I guess the bigger question is what does an online retail business use facial recognition for? Or maybe even better is how do they use facial recognition software?

Perhaps the very ones here wanting to protest Amazon for not providing facial recognition software, in years past would have decried its use as being Orwellian. And now they are protesting for its use.

See how easily the masses are manipulated.

I don't understand this logic. "Misidentify people of color"; if the person committed a crime (looting) then they need to be punished...brown, black, white, yellow, blue, red, purple -- skin color does not matter. I'm confused.

~Sovereign

@HiddenI

Yes, yes. We've seen just that the past few weeks.

a reply to: Bluntone22

The idea of facial recognition being used by anyone doesn't sit well with me, but I can see the value in it as a possible lead in crime investigation.

I don't think a conviction hinging on a *match* should be permitted though.

a reply to: puzzled2

No, like it says in the opening post, it was the US government that found out that facial recognition software (not just Amazon's) have a tendency of misidentifying non-white subjects.

That was the result of a government study of 189 software algorithms from 99 developers, that showed that:

False positives: Using the higher quality Application photos, false positive rates are highest in West and East African and East Asian people, and lowest in Eastern European individuals. This effect is generally large, with a factor of 100 more false positives between countries. However, with a number of algorithms developed in China this effect is reversed, with low false positive rates on East Asian faces. With domestic law enforcement images, the highest false positives are in American Indians, with elevated rates in African American and Asian populations; the relative ordering depends on sex and varies with algorithm.

We found false positives to be higher in women than men, and this is consistent across algorithms and datasets. This effect is smaller than that due to race.

We found elevated false positives in the elderly and in children; the effects were larger in
the oldest and youngest, and smallest in middle-aged adults.

Source

a reply to: Mandroid7

Apparently, police in the US can ask for access to the Ring (an Amazon subsidiary) surveillance cameras and uses facial recognition algorithms to help find "people of interest" in their cases.

Through the portal, law enforcement are also able to enlist the help of the community on active investigations by requesting footage from users through the Video Request tool. With the tool, local law enforcement can ask Ring to request video footage from device owners who are in the area of an active investigation. When making a video request to Ring, law enforcement must reference a relevant case, and can only request video recordings within a limited time and area. With each request, customers decide whether to share all relevant videos, review and select certain videos to share, take no action (decline), or opt-out of all future requests.

Source

Perhaps it should be "amid questions about the company's commitment to be(or not be) a snitch"

a reply to: Ahabstar

Amazon has stopped being mostly an online retail business long ago, a large part of their revenue now comes from the Amazon Web Services, online software services people can use in a "pay as you go" method in many ways.

Those services go from basic storage to AI services that users can use as they want.