Twitter's algorithm would be racist and cut out photos in a weird way.
The world's largest microblogging service, Twitter, admits flaws in its automatic image cropping and is now investigating the machine learning algorithm behind it.
According to the US company, the selection of a section should in future be made “by the people who use our products”. The systems are still too error-prone, as they say. Users had previously complained that the algorithm used to select images was biased and preferred faces of white people.
“Damage of representation”
In the course of the review, Twitter has now come to the conclusion that the algorithm, for example, prefers men over women in addition to skin color when selecting images. However, Twitter does not accept the accusation that women's chest areas or legs are preferred when cropping images. Instead of focusing on physical characteristics, the software focuses on the numbers on sports jerseys when selecting images.
“Even if the saliency algorithm were adjusted to reflect perfect equality between ethnicity and gender, we are concerned about the representational harm of the automated algorithm if people are not allowed to represent themselves on the platform as they wish,” shares Twitter. Salience also creates other possible disadvantages. The company explicitly mentions a lack of tact when it comes to cultural differences. This is now being revised.
[mk_ad]
Source: press release
Also interesting:
After another woman was murdered by her (ex) partner in Vienna, many people expressed their anger and frustration on Twitter.
Notes:
1) This content reflects the current state of affairs at the time of publication. The reproduction of individual images, screenshots, embeds or video sequences serves to discuss the topic. 2) Individual contributions were created through the use of machine assistance and were carefully checked by the Mimikama editorial team before publication. ( Reason )