In the visual tweets shared on Twitter, one of the most active social media channels in the world; If there are two fair-skinned and dark-skinned people, the algorithm highlights the white-skinned person. The selections of the Twitter algorithm, which lists the visual posts with their prominent features in the stream, fueled the racism agenda on the platform.
If you are a die-hard Twitter user, you may have noticed that tweets with multiple images appear as a small gallery in the stream. In this gallery view, up to 4 of the visuals in the shares are arranged in a reduced size, and the focus point of the visuals is placed on the thumbnails.
For example, if there is a person on the far right of a panoramic landscape photo, Twitter will show you the rightmost person, not the middle view section in the small gallery view. When you click it, the real size version of the photo is uploaded. Twitter’s artificial intelligence is responsible for this automation process. The person, text and smiling face in the image are selected and highlighted in the small picture.
However, Twitter’s artificial intelligence sparked racism debates with the elections it made. Try opening photos from these posts that seem completely normal at first glance:
The first post includes US politicians Mitch McConnell and Barrack Obama. One of these people is light and the other is dark skinned. However, in the thumbnail gallery view, only the white-skinned Mitch McConnell is highlighted, and when we click, we see that Barrack Obama is also in the image. Moreover, even if the location of Obama and McConnell changes in the later stages of the experiment, the result does not change, Twitter’s algorithm continues to insist.
In the second post, the colors of the tie are changed and the photograph of the white-skinned person is still highlighted, even if there is shifting in the small picture:
When the images are made negative, that is, when the colors are reversed, the result changes: