Tampere
28 Mar, Thursday
7° C

The library of essays of Proakatemia

Are we Automating Racism?



Kirjoittanut: Omar Puebla Roldan - tiimistä FLIP Solutions.

Esseen tyyppi: Blogiessee / 1 esseepistettä.
Esseen arvioitu lukuaika on 2 minuuttia.

Are we Automating Racism?

 

Twitter is investigating users who discovered its picture-cropping algorithm prefers white faces to black ones. Some users on Twitter noticed when two photos, one of a black face and the other with a white face, in the same post, Twitter often showed only the white face on mobile.

 

Time of experiment, many users on Twitter experimented using one picture that  Barack Obama and Mitch McConnell, the result was the 90%  of the post-Tiwteter precut image chose Mitch McConnell.

More users on Twitter they noticed this too, and they started to do more experiments. Some of them were more creative than others: White Michael Jackson and Black Michael Jackson, Back Men photoshopped as a White man, Golden retriever and Black retriever, Simpsons characters Carl and Lenny, and some other cartoons where are white characters and black characters.

Many of the results show that the algorithm chose white colour people, but this is only one social media problem. For example, many of the beauty filters on Tiktok reduce people’s noses; many Black African users complain about this, the concept of having a small nose synonymous with beauty.

 

Zoom has crappy face detection, a black African American tried the background feature for this zoom meeting, and he noticed the algorithm cut his face many times during the virtual meeting, on the other hand, nothing happened to the White person.

 

Instagram is biased against women, especially Black women and minorities, especially fat Black women. When a body is not understood, I think the algorithm goes against what it’s taught as the norm, which is white, slim women as the ideal in the media.

 

Are we Automating Racism?

Is this just a commentary problem?

I found that many social media algorithms were created ten years ago, and the people who made them didn’t consider using more images containing black people. They did their research, they did their analysis, but to who has reported those results?

 

When the Ai came into our lives, we thought it would be neutral that robots would not discriminate our skin color, or religion, or gender.

 

We will face this and more problems in the future, is only white people creating algorithms? Who is deciding this use those algorithms?

 

 

Twitter apologises for ‘racist’ image-cropping algorithm

https://www.theguardian.com/technology/2020/sep/21/twitter-apologises-for-racist-image-cropping-algorithm

 

Twitter finds racial bias in image-cropping AI

https://www.bbc.com/news/technology-57192898

Post a Comment

Add Comment
Loading...

Cancel
Viewing Highlight
Loading...
Highlight
Close