Twitter has finally found an answer to a technical problem discovered by the platform’s research team: the “saliency algorithm,” which decides how images in Twitter previews are cropped before they are clicked in full size.<\/p>\n\n\n\n
According to the Twitter research team, the algorithm tends to cut out black faces in favor of white faces.<\/p>\n\n\n\n
The pattern applied to images of former U.S. President Barack Obama and Senator Mitch McConnell – and stock images of businessmen of different ethnicities.<\/p>\n\n\n\n
Bogdan Kulynyc, a graduate student at the Swiss Federal Institute of Technology, explained that the “saliency” of a face in an image can be increased, thereby preventing the likelihood of the image from being obscured by the cropping algorithm by “making the person’s skin lighter or warmer and smoother: and quite often changing the appearance to that of a younger, more slim, and more stereotypically feminine person.”<\/p>\n\n\n\n
This solution won Kulynyc first prize in a Twitter organized competition, and, as Twitter stated, Kulynyc’s discovery showed that beauty filters can be used to manipulate the algorithm and show “how algorithmic models amplify real-world biases and societal expectations of beauty.”<\/p>\n\n\n