Twitter's automatic image cropping is prejudiced against whites and women.
Twitter promised to re-evaluate its image cropping algorithm after users complained of slotxo bias. On Wednesday, the company released the results of the study. When people began documenting problems with the algorithm, Twitter said the initial investigation showed no evidence of racial or gender bias. Today's research paints a different picture.In comparing the algorithm's preferences between men and women in addition to treating whites and blacks, the company found that women liked women 8 percent differently and liked whites by 4 percent.
Arithm reflected the gaze of the man in him. Testing with images of Twitter women found that the algorithm cuts only three images of every 100 images to a location other than the face, and when doing so, it doesn't focus on any part of the body, but limits its composition. Narrowed things such as numbers on sports jerseys The company said one possible reason for its algorithm's cropping options could be the demands of high-contrast images. But noted that it was not an excuse. "Machine learning cropping is fundamentally flawed, as it removes the user agency and restricts
the user's expression of their identity and self-worth, rather than the normative gaze on which parts of the user agency work. "One of our conclusions is that not everything on Twitter is a good choice for algorithms, and in this case, the image crop method is the best decision made by the people who use it." people"As a result, the company recently released full-sized image previews on Android and iOS and said it plans to implement further changes to how the platform handles media in the future.